eval-harness

arms

Arms and configs

686 arms from 267 configs on 16 datasets. An arm is one config on one dataset scored one way: measured, rescored against another reference, judged, or run again (the how column says which; group by config to see them together). Pick arms on one dataset to open them in the inspector.

all time, 2026-09-25 to 2026-11-02

01every arm

group by

only

686 shown
armhow measureddatasetmodel · providerdeciding measurerepeatscost / itemlatencyruns in periodspend in period
ag_anthropic_l_n200_v1measuredag_news_200eval-opus-5
litellm
correct 0.9001$0.00091965 ms1—
ag_anthropic_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.900 → 0.985 (+0.085) vs measured
ag_news_200eval-opus-5
litellm
correct 0.9851$0.00091965 ms0—
ag_anthropic_m_n200_v1measuredag_news_200eval-claude-sonnet
litellm
correct 0.9101$0.00041916 ms1—
ag_anthropic_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.910 → 0.965 (+0.055) vs measured
ag_news_200eval-claude-sonnet
litellm
correct 0.9651$0.00041916 ms0—
ag_anthropic_s_n200_v1measuredag_news_200eval-claude-haiku
litellm
correct 0.8401$0.0001987 ms1—
ag_anthropic_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.840 → 0.925 (+0.085) vs measured
ag_news_200eval-claude-haiku
litellm
correct 0.9251$0.0001987 ms0—
ag_bart_mnli_n200_v1measuredag_news_200facebook/bart-large-mnli
hf_zeroshot · local
correct 0.7001—1387 ms1—
ag_bart_mnli_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.700 → 0.755 (+0.055) vs measured
ag_news_200facebook/bart-large-mnli
hf_zeroshot · local
correct 0.7551—1387 ms0—
ag_bert_base_fy_n200_v1measuredag_news_200fabriceyhc/bert-base-uncased-ag_news
hf_local · local
correct 0.9501—90 ms1—
ag_bert_base_fy_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.950 → 0.905 (-0.045) vs measured
ag_news_200fabriceyhc/bert-base-uncased-ag_news
hf_local · local
correct 0.9051—90 ms0—
ag_bert_base_fy_n200_v1 · linux
retired in r2
run again (linux)
0.950 → 0.950 (+0.000) vs measured
ag_news_200fabriceyhc/bert-base-uncased-ag_news
hf_local · local
correct 0.9501—202 ms1—
ag_bert_base_ta_n200_v1
leader
measuredag_news_200textattack/bert-base-uncased-ag-news
hf_local · local
correct 0.9601—76 ms1—
ag_bert_base_ta_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.960 → 0.905 (-0.055) vs measured
ag_news_200textattack/bert-base-uncased-ag-news
hf_local · local
correct 0.9051—76 ms0—
ag_bert_base_ta_n200_v1 · linux
retired in r2
run again (linux)
0.960 → 0.960 (+0.000) vs measured
ag_news_200textattack/bert-base-uncased-ag-news
hf_local · local
correct 0.9601—57 ms1—
ag_bert_mini_n200_v1
baseline
measuredag_news_200mrm8488/bert-mini-finetuned-age_news-classification
hf_local · local
correct 0.9451—7 ms1—
ag_bert_mini_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.945 → 0.905 (-0.040) vs measured
ag_news_200mrm8488/bert-mini-finetuned-age_news-classification
hf_local · local
correct 0.9051—7 ms0—
ag_bert_mini_n200_v1 · linux
retired in r2
run again (linux)
0.945 → 0.945 (+0.000) vs measured
ag_news_200mrm8488/bert-mini-finetuned-age_news-classification
hf_local · local
correct 0.9451—21 ms1—
ag_constant_n200_v1measuredag_news_200—
constant · local
correct 0.2501—0 ms1—
ag_constant_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.250 → 0.225 (-0.025) vs measured
ag_news_200—
constant · local
correct 0.2251—0 ms0—
ag_deepseek_l_n200_v1measuredag_news_200eval-deepseek-pro
litellm
correct 0.8701$0.00011323 ms1—
ag_deepseek_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.870 → 0.955 (+0.085) vs measured
ag_news_200eval-deepseek-pro
litellm
correct 0.9551$0.00011323 ms0—
ag_deepseek_m_n200_v1measuredag_news_200eval-deepseek-41flash
litellm
correct 0.8451$0.0000890 ms1—
ag_deepseek_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.845 → 0.925 (+0.080) vs measured
ag_news_200eval-deepseek-41flash
litellm
correct 0.9251$0.0000890 ms0—
ag_deepseek_s_n200_v1measuredag_news_200eval-deepseek-flash
litellm
correct 0.8401$0.00001602 ms1—
ag_deepseek_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.840 → 0.925 (+0.085) vs measured
ag_news_200eval-deepseek-flash
litellm
correct 0.9251$0.00001602 ms0—
ag_gemma_l_n200_v1measuredag_news_200eval-gemma-31b
litellm
correct 0.8701$0.0000911 ms1—
ag_gemma_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.870 → 0.955 (+0.085) vs measured
ag_news_200eval-gemma-31b
litellm
correct 0.9551$0.0000911 ms0—
ag_gemma_m_n200_v1measuredag_news_200eval-gemma-26b
litellm
correct 0.8551$0.0000666 ms1—
ag_gemma_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.855 → 0.940 (+0.085) vs measured
ag_news_200eval-gemma-26b
litellm
correct 0.9401$0.0000666 ms0—
ag_gemma_s_n200_v1measuredag_news_200eval-gemma-3-27b
litellm
correct 0.8751$0.0000664 ms1—
ag_gemma_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.875 → 0.960 (+0.085) vs measured
ag_news_200eval-gemma-3-27b
litellm
correct 0.9601$0.0000664 ms0—
ag_glm_l_n200_v1measuredag_news_200eval-glm-5
litellm
correct 0.8601$0.00011871 ms1—
ag_glm_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.860 → 0.945 (+0.085) vs measured
ag_news_200eval-glm-5
litellm
correct 0.9451$0.00011871 ms0—
ag_glm_m_n200_v1measuredag_news_200eval-glm-46
litellm
correct 0.8601$0.00012262 ms1—
ag_glm_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.860 → 0.945 (+0.085) vs measured
ag_news_200eval-glm-46
litellm
correct 0.9451$0.00012262 ms0—
ag_glm_s_n200_v1measuredag_news_200eval-glm-45-air
litellm
correct 0.8501$0.00001154 ms1—
ag_glm_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.850 → 0.935 (+0.085) vs measured
ag_news_200eval-glm-45-air
litellm
correct 0.9351$0.00001154 ms0—
ag_keyword_n200_v1measuredag_news_200—
keyword · local
correct 0.6701—0 ms1—
ag_keyword_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.670 → 0.670 (+0.000) vs measured
ag_news_200—
keyword · local
correct 0.6701—0 ms0—
ag_llama_l_n200_v1measuredag_news_200eval-llama-maverick
litellm
correct 0.8401$0.0000804 ms1—
ag_llama_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.840 → 0.925 (+0.085) vs measured
ag_news_200eval-llama-maverick
litellm
correct 0.9251$0.0000804 ms0—
ag_llama_m_n200_v1measuredag_news_200eval-llama-70b
litellm
correct 0.8801$0.00001193 ms1—
ag_llama_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.880 → 0.965 (+0.085) vs measured
ag_news_200eval-llama-70b
litellm
correct 0.9651$0.00001193 ms0—
ag_llama_s_n200_v1measuredag_news_200eval-llama-scout
litellm
correct 0.8351$0.0000784 ms1—
ag_llama_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.835 → 0.920 (+0.085) vs measured
ag_news_200eval-llama-scout
litellm
correct 0.9201$0.0000784 ms0—
ag_mistral_l_n200_v1measuredag_news_200eval-mistral-large
litellm
correct 0.8751$0.00017662 ms1—
ag_mistral_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.875 → 0.950 (+0.075) vs measured
ag_news_200eval-mistral-large
litellm
correct 0.9501$0.00017662 ms0—
ag_mistral_m_n200_v1measuredag_news_200eval-mistral-medium
litellm
correct 0.8601$0.0000437 ms1—
ag_mistral_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.860 → 0.945 (+0.085) vs measured
ag_news_200eval-mistral-medium
litellm
correct 0.9451$0.0000437 ms0—
ag_mistral_s_n200_v1measuredag_news_200eval-mistral-small
litellm
correct 0.8601$0.0000732 ms1—
ag_mistral_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.860 → 0.930 (+0.070) vs measured
ag_news_200eval-mistral-small
litellm
correct 0.9301$0.0000732 ms0—
ag_openai_l_n200_v1measuredag_news_200eval-gpt-6sol
litellm
correct 0.8901$0.00021207 ms1—
ag_openai_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.890 → 0.975 (+0.085) vs measured
ag_news_200eval-gpt-6sol
litellm
correct 0.9751$0.00021207 ms0—
ag_openai_m_n200_v1measuredag_news_200eval-gpt-55
litellm
correct 0.8951$0.00061105 ms1—
ag_openai_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.895 → 0.980 (+0.085) vs measured
ag_news_200eval-gpt-55
litellm
correct 0.9801$0.00061105 ms0—
ag_openai_s_n200_v1measuredag_news_200eval-gpt-mini
litellm
correct 0.8601$0.0001859 ms1—
ag_openai_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.860 → 0.945 (+0.085) vs measured
ag_news_200eval-gpt-mini
litellm
correct 0.9451$0.0001859 ms0—
ag_qwen_l_n200_v1measuredag_news_200eval-qwen-3max
litellm
correct 0.8701$0.00012849 ms1—
ag_qwen_l_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.870 → 0.955 (+0.085) vs measured
ag_news_200eval-qwen-3max
litellm
correct 0.9551$0.00012849 ms0—
ag_qwen_m_n200_v1measuredag_news_200eval-qwen-37plus
litellm
correct 0.8851$0.0000911 ms1—
ag_qwen_m_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.885 → 0.960 (+0.075) vs measured
ag_news_200eval-qwen-37plus
litellm
correct 0.9601$0.0000911 ms0—
ag_qwen_s_n200_v1measuredag_news_200eval-qwen-flash
litellm
correct 0.8751$0.00001144 ms1—
ag_qwen_s_n200_v1 · adjudicatedrescored against gold/ag_news_200_adj1
0.875 → 0.960 (+0.085) vs measured
ag_news_200eval-qwen-flash
litellm
correct 0.9601$0.00001144 ms0—
cnn_bart_l_v1measuredcnn_dailymail_20facebook/bart-large-cnn
hf_local · local
coverage 0.3661—11861 ms1—
cnn_lead3_v1measuredcnn_dailymail_20—
lead_k · local
coverage 0.2621—0 ms1—
cnn_anthropic_l_n200_v1measuredcnn_dailymail_200eval-opus-5
litellm
coverage 0.3151$0.00983402 ms1—
cnn_anthropic_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-opus-5
litellm
judge_won 0.4021$0.00983399 ms0—
cnn_anthropic_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-opus-5
litellm
judge_score 4.9761$0.00983402 ms0—
cnn_anthropic_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-opus-5
litellm
judge_score 4.8881$0.00983402 ms0—
cnn_anthropic_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.315 → 1.000 (+0.685) vs measured
cnn_dailymail_200eval-opus-5
litellm
coverage 1.0001$0.00983402 ms0—
cnn_anthropic_m_n200_v1measuredcnn_dailymail_200eval-claude-sonnet
litellm
coverage 0.3241$0.00373034 ms1—
cnn_anthropic_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-claude-sonnet
litellm
judge_won 0.1441$0.00373033 ms0—
cnn_anthropic_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-claude-sonnet
litellm
judge_score 4.9631$0.00373032 ms0—
cnn_anthropic_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-claude-sonnet
litellm
judge_score 4.8091$0.00373034 ms0—
cnn_anthropic_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.324 → 0.491 (+0.168) vs measured
cnn_dailymail_200eval-claude-sonnet
litellm
coverage 0.4911$0.00373034 ms0—
cnn_anthropic_s_n200_v1measuredcnn_dailymail_200eval-claude-haiku
litellm
coverage 0.3241$0.00132025 ms1—
cnn_anthropic_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-claude-haiku
litellm
judge_won 0.2261$0.00132023 ms0—
cnn_anthropic_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-claude-haiku
litellm
judge_score 4.9521$0.00132020 ms0—
cnn_anthropic_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-claude-haiku
litellm
judge_score 4.7281$0.00132025 ms0—
cnn_anthropic_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.324 → 0.469 (+0.145) vs measured
cnn_dailymail_200eval-claude-haiku
litellm
coverage 0.4691$0.00132025 ms0—
cnn_bart_l_n200_v1
baselineleaderrecommended
measuredcnn_dailymail_200facebook/bart-large-cnn
hf_local · local
coverage 0.3461—11010 ms1—
cnn_bart_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200facebook/bart-large-cnn
hf_local · local
judge_won 0.0001—11007 ms0—
cnn_bart_l_n200_v1 · judged/judge-summ-pointwise-v1judged by judge-summ-pointwise-v1cnn_dailymail_200facebook/bart-large-cnn
hf_local · local
judge_score_unvalidated 4.1811—11010 ms0—
cnn_bart_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200facebook/bart-large-cnn
hf_local · local
judge_score 3.8681—10996 ms0—
cnn_bart_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200facebook/bart-large-cnn
hf_local · local
judge_score 3.8211—11010 ms0—
cnn_bart_l_n200_v1 · linux
retired in r2
run again (linux)
0.346 → 0.346 (+0.000) vs measured
cnn_dailymail_200facebook/bart-large-cnn
hf_local · local
coverage 0.3461—6507 ms1—
cnn_bart_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.346 → 0.317 (-0.029) vs measured
cnn_dailymail_200facebook/bart-large-cnn
hf_local · local
coverage 0.3171—11010 ms0—
cnn_bart_l_xsum_n200_v1measuredcnn_dailymail_200facebook/bart-large-xsum
hf_local · local
coverage 0.2071—4621 ms1—
cnn_bart_l_xsum_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200facebook/bart-large-xsum
hf_local · local
judge_won 0.0001—4609 ms0—
cnn_bart_l_xsum_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200facebook/bart-large-xsum
hf_local · local
judge_score 3.5161—4605 ms0—
cnn_bart_l_xsum_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200facebook/bart-large-xsum
hf_local · local
judge_score 3.3661—4621 ms0—
cnn_bart_l_xsum_n200_v1 · linux
retired in r2
run again (linux)
0.207 → 0.207 (+0.000) vs measured
cnn_dailymail_200facebook/bart-large-xsum
hf_local · local
coverage 0.2071—4248 ms1—
cnn_bart_l_xsum_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.207 → 0.137 (-0.070) vs measured
cnn_dailymail_200facebook/bart-large-xsum
hf_local · local
coverage 0.1371—4621 ms0—
cnn_bart_m_n200_v1measuredcnn_dailymail_200sshleifer/distilbart-cnn-12-6
hf_local · local
coverage 0.3371—5062 ms1—
cnn_bart_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200sshleifer/distilbart-cnn-12-6
hf_local · local
judge_won 0.0001—5048 ms0—
cnn_bart_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200sshleifer/distilbart-cnn-12-6
hf_local · local
judge_score 3.5641—5054 ms0—
cnn_bart_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200sshleifer/distilbart-cnn-12-6
hf_local · local
judge_score 3.2351—5062 ms0—
cnn_bart_m_n200_v1 · linux
retired in r2
run again (linux)
0.337 → 0.337 (+0.000) vs measured
cnn_dailymail_200sshleifer/distilbart-cnn-12-6
hf_local · local
coverage 0.3371—4375 ms1—
cnn_bart_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.337 → 0.320 (-0.017) vs measured
cnn_dailymail_200sshleifer/distilbart-cnn-12-6
hf_local · local
coverage 0.3201—5062 ms0—
cnn_bart_s_n200_v1measuredcnn_dailymail_200sshleifer/distilbart-cnn-6-6
hf_local · local
coverage 0.3281—4484 ms1—
cnn_bart_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200sshleifer/distilbart-cnn-6-6
hf_local · local
judge_won 0.0001—4485 ms0—
cnn_bart_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200sshleifer/distilbart-cnn-6-6
hf_local · local
judge_score 3.1851—4489 ms0—
cnn_bart_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200sshleifer/distilbart-cnn-6-6
hf_local · local
judge_score 2.9461—4484 ms0—
cnn_bart_s_n200_v1 · linux
retired in r2
run again (linux)
0.328 → 0.328 (+0.000) vs measured
cnn_dailymail_200sshleifer/distilbart-cnn-6-6
hf_local · local
coverage 0.3281—3420 ms1—
cnn_bart_s_n200_v1 · repeats
retired in r3
run again (repeats)
0.328 → 0.328 (+0.000) vs measured
cnn_dailymail_200sshleifer/distilbart-cnn-6-6
hf_local · local
coverage 0.3281—3766 ms1—
cnn_bart_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.328 → 0.306 (-0.022) vs measured
cnn_dailymail_200sshleifer/distilbart-cnn-6-6
hf_local · local
coverage 0.3061—4484 ms0—
cnn_deepseek_l_n200_v1measuredcnn_dailymail_200eval-deepseek-pro
litellm
coverage 0.3321$0.00072665 ms1—
cnn_deepseek_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-deepseek-pro
litellm
judge_won 0.0141$0.00072659 ms0—
cnn_deepseek_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-deepseek-pro
litellm
judge_score 4.8771$0.00072666 ms0—
cnn_deepseek_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-deepseek-pro
litellm
judge_score 4.7631$0.00072665 ms0—
cnn_deepseek_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.332 → 0.399 (+0.066) vs measured
cnn_dailymail_200eval-deepseek-pro
litellm
coverage 0.3991$0.00072665 ms0—
cnn_deepseek_m_n200_v1measuredcnn_dailymail_200eval-deepseek-41flash
litellm
coverage 0.3461$0.00021595 ms1—
cnn_deepseek_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-deepseek-41flash
litellm
judge_won 0.0831$0.00021592 ms0—
cnn_deepseek_m_n200_v1 · judged/judge-summ-pointwise-v1judged by judge-summ-pointwise-v1cnn_dailymail_200eval-deepseek-41flash
litellm
judge_score_unvalidated 4.9401$0.00021595 ms0—
cnn_deepseek_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-deepseek-41flash
litellm
judge_score 4.9231$0.00021589 ms0—
cnn_deepseek_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-deepseek-41flash
litellm
judge_score 4.8541$0.00021595 ms0—
cnn_deepseek_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.346 → 0.471 (+0.125) vs measured
cnn_dailymail_200eval-deepseek-41flash
litellm
coverage 0.4711$0.00021595 ms0—
cnn_deepseek_s_n200_v1measuredcnn_dailymail_200eval-deepseek-flash
litellm
coverage 0.3281$0.00012753 ms1—
cnn_deepseek_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-deepseek-flash
litellm
judge_won 0.0041$0.00012714 ms0—
cnn_deepseek_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-deepseek-flash
litellm
judge_score 4.8181$0.00012753 ms0—
cnn_deepseek_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-deepseek-flash
litellm
judge_score 4.6441$0.00012753 ms0—
cnn_deepseek_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.328 → 0.388 (+0.059) vs measured
cnn_dailymail_200eval-deepseek-flash
litellm
coverage 0.3881$0.00012753 ms0—
cnn_gemma_l_n200_v1measuredcnn_dailymail_200eval-gemma-31b
litellm
coverage 0.3201$0.00012591 ms1—
cnn_gemma_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-gemma-31b
litellm
judge_won 0.0011$0.00012600 ms0—
cnn_gemma_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-gemma-31b
litellm
judge_score 4.7151$0.00012601 ms0—
cnn_gemma_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-gemma-31b
litellm
judge_score 4.6031$0.00012591 ms0—
cnn_gemma_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.320 → 0.342 (+0.021) vs measured
cnn_dailymail_200eval-gemma-31b
litellm
coverage 0.3421$0.00012591 ms0—
cnn_gemma_m_n200_v1measuredcnn_dailymail_200eval-gemma-26b
litellm
coverage 0.3141$0.00011873 ms1—
cnn_gemma_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-gemma-26b
litellm
judge_won 0.0051$0.00011871 ms0—
cnn_gemma_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-gemma-26b
litellm
judge_score 4.7071$0.00011873 ms0—
cnn_gemma_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-gemma-26b
litellm
judge_score 4.5851$0.00011873 ms0—
cnn_gemma_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.314 → 0.342 (+0.029) vs measured
cnn_dailymail_200eval-gemma-26b
litellm
coverage 0.3421$0.00011873 ms0—
cnn_gemma_s_n200_v1measuredcnn_dailymail_200eval-gemma-3-27b
litellm
coverage 0.3121$0.00016956 ms1—
cnn_gemma_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-gemma-3-27b
litellm
judge_won 0.0021$0.00016932 ms0—
cnn_gemma_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-gemma-3-27b
litellm
judge_score 4.8011$0.00016989 ms0—
cnn_gemma_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-gemma-3-27b
litellm
judge_score 4.4291$0.00016956 ms0—
cnn_gemma_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.312 → 0.350 (+0.039) vs measured
cnn_dailymail_200eval-gemma-3-27b
litellm
coverage 0.3501$0.00016956 ms0—
cnn_glm_l_n200_v1measuredcnn_dailymail_200eval-glm-5
litellm
coverage 0.3121$0.00085689 ms1—
cnn_glm_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-glm-5
litellm
judge_won 0.0061$0.00085718 ms0—
cnn_glm_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-glm-5
litellm
judge_score 4.6071$0.00085739 ms0—
cnn_glm_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-glm-5
litellm
judge_score 4.3841$0.00085689 ms0—
cnn_glm_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.312 → 0.375 (+0.063) vs measured
cnn_dailymail_200eval-glm-5
litellm
coverage 0.3751$0.00085689 ms0—
cnn_glm_m_n200_v1measuredcnn_dailymail_200eval-glm-46
litellm
coverage 0.3321$0.00053230 ms1—
cnn_glm_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-glm-46
litellm
judge_won 0.0051$0.00053243 ms0—
cnn_glm_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-glm-46
litellm
judge_score 4.8611$0.00053234 ms0—
cnn_glm_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-glm-46
litellm
judge_score 4.7181$0.00053230 ms0—
cnn_glm_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.332 → 0.389 (+0.057) vs measured
cnn_dailymail_200eval-glm-46
litellm
coverage 0.3891$0.00053230 ms0—
cnn_glm_s_n200_v1measuredcnn_dailymail_200eval-glm-45-air
litellm
coverage 0.3141$0.00021610 ms1—
cnn_glm_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-glm-45-air
litellm
judge_won 0.0011$0.00021612 ms0—
cnn_glm_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-glm-45-air
litellm
judge_score 4.5301$0.00021615 ms0—
cnn_glm_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-glm-45-air
litellm
judge_score 4.4201$0.00021610 ms0—
cnn_glm_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.314 → 0.303 (-0.011) vs measured
cnn_dailymail_200eval-glm-45-air
litellm
coverage 0.3031$0.00021610 ms0—
cnn_lead3_n200_v1measuredcnn_dailymail_200—
lead_k · local
coverage 0.2721—0 ms1—
cnn_lead3_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200—
lead_k · local
judge_won 0.0001—0 ms0—
cnn_lead3_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200—
lead_k · local
judge_score 3.4211—0 ms0—
cnn_lead3_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200—
lead_k · local
judge_score 3.3391—0 ms0—
cnn_lead3_n200_v1 · linux
retired in r2
run again (linux)
0.272 → 0.272 (+0.000) vs measured
cnn_dailymail_200—
lead_k · local
coverage 0.2721—0 ms1—
cnn_lead3_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.272 → 0.300 (+0.029) vs measured
cnn_dailymail_200—
lead_k · local
coverage 0.3001—0 ms0—
cnn_llama_l_n200_v1measuredcnn_dailymail_200eval-llama-maverick
litellm
coverage 0.3391$0.00023473 ms1—
cnn_llama_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-llama-maverick
litellm
judge_won 0.0031$0.00023474 ms0—
cnn_llama_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-llama-maverick
litellm
judge_score 4.7001$0.00023495 ms0—
cnn_llama_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-llama-maverick
litellm
judge_score 4.5291$0.00023473 ms0—
cnn_llama_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.339 → 0.370 (+0.031) vs measured
cnn_dailymail_200eval-llama-maverick
litellm
coverage 0.3701$0.00023473 ms0—
cnn_llama_m_n200_v1measuredcnn_dailymail_200eval-llama-70b
litellm
coverage 0.3381$0.00024045 ms1—
cnn_llama_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-llama-70b
litellm
judge_won 0.0011$0.00024066 ms0—
cnn_llama_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-llama-70b
litellm
judge_score 4.6741$0.00024072 ms0—
cnn_llama_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-llama-70b
litellm
judge_score 4.4881$0.00024045 ms0—
cnn_llama_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.338 → 0.358 (+0.019) vs measured
cnn_dailymail_200eval-llama-70b
litellm
coverage 0.3581$0.00024045 ms0—
cnn_llama_s_n200_v1measuredcnn_dailymail_200eval-llama-scout
litellm
coverage 0.3331$0.00012388 ms1—
cnn_llama_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-llama-scout
litellm
judge_won 0.0011$0.00012370 ms0—
cnn_llama_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-llama-scout
litellm
judge_score 4.7041$0.00012392 ms0—
cnn_llama_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-llama-scout
litellm
judge_score 4.4691$0.00012388 ms0—
cnn_llama_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.333 → 0.388 (+0.055) vs measured
cnn_dailymail_200eval-llama-scout
litellm
coverage 0.3881$0.00012388 ms0—
cnn_mistral_l_n200_v1measuredcnn_dailymail_200eval-mistral-large
litellm
coverage 0.3081$0.000511045 ms1—
cnn_mistral_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-mistral-large
litellm
judge_won 0.0031$0.000510529 ms0—
cnn_mistral_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-mistral-large
litellm
judge_score 4.6981$0.000511180 ms0—
cnn_mistral_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-mistral-large
litellm
judge_score 4.0371$0.000511045 ms0—
cnn_mistral_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.308 → 0.337 (+0.029) vs measured
cnn_dailymail_200eval-mistral-large
litellm
coverage 0.3371$0.000511045 ms0—
cnn_mistral_m_n200_v1measuredcnn_dailymail_200eval-mistral-medium
litellm
coverage 0.3041$0.00042654 ms1—
cnn_mistral_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-mistral-medium
litellm
judge_won 0.0001$0.00042648 ms0—
cnn_mistral_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-mistral-medium
litellm
judge_score 4.7081$0.00042659 ms0—
cnn_mistral_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-mistral-medium
litellm
judge_score 4.3091$0.00042654 ms0—
cnn_mistral_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.304 → 0.320 (+0.016) vs measured
cnn_dailymail_200eval-mistral-medium
litellm
coverage 0.3201$0.00042654 ms0—
cnn_mistral_s_n200_v1measuredcnn_dailymail_200eval-mistral-small
litellm
coverage 0.3261$0.00012820 ms1—
cnn_mistral_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-mistral-small
litellm
judge_won 0.0041$0.00012810 ms0—
cnn_mistral_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-mistral-small
litellm
judge_score 4.7981$0.00012826 ms0—
cnn_mistral_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-mistral-small
litellm
judge_score 4.4081$0.00012820 ms0—
cnn_mistral_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.326 → 0.382 (+0.057) vs measured
cnn_dailymail_200eval-mistral-small
litellm
coverage 0.3821$0.00012820 ms0—
cnn_openai_l_n200_v1measuredcnn_dailymail_200eval-gpt-6sol
litellm
coverage 0.3271$0.00221775 ms1—
cnn_openai_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-gpt-6sol
litellm
judge_won 0.0041$0.00221775 ms0—
cnn_openai_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-gpt-6sol
litellm
judge_score 4.7731$0.00221769 ms0—
cnn_openai_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-gpt-6sol
litellm
judge_score 4.8471$0.00221775 ms0—
cnn_openai_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.327 → 0.371 (+0.044) vs measured
cnn_dailymail_200eval-gpt-6sol
litellm
coverage 0.3711$0.00221775 ms0—
cnn_openai_m_n200_v1measuredcnn_dailymail_200eval-gpt-55
litellm
coverage 0.3271$0.00541745 ms1—
cnn_openai_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-gpt-55
litellm
judge_won 0.0041$0.00531746 ms0—
cnn_openai_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-gpt-55
litellm
judge_score 4.7951$0.00531746 ms0—
cnn_openai_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-gpt-55
litellm
judge_score 4.8111$0.00541745 ms0—
cnn_openai_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.327 → 0.359 (+0.033) vs measured
cnn_dailymail_200eval-gpt-55
litellm
coverage 0.3591$0.00541745 ms0—
cnn_openai_s_n200_v1measuredcnn_dailymail_200eval-gpt-mini
litellm
coverage 0.3171$0.00081277 ms1—
cnn_openai_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-gpt-mini
litellm
judge_won 0.0081$0.00081277 ms0—
cnn_openai_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-gpt-mini
litellm
judge_score 4.8251$0.00081277 ms0—
cnn_openai_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-gpt-mini
litellm
judge_score 4.7471$0.00081277 ms0—
cnn_openai_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.317 → 0.384 (+0.067) vs measured
cnn_dailymail_200eval-gpt-mini
litellm
coverage 0.3841$0.00081277 ms0—
cnn_qwen_l_n200_v1measuredcnn_dailymail_200eval-qwen-3max
litellm
coverage 0.3111$0.00082945 ms1—
cnn_qwen_l_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-qwen-3max
litellm
judge_won 0.0591$0.00082943 ms0—
cnn_qwen_l_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-qwen-3max
litellm
judge_score 4.8121$0.00082946 ms0—
cnn_qwen_l_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-qwen-3max
litellm
judge_score 4.5731$0.00082945 ms0—
cnn_qwen_l_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.311 → 0.398 (+0.087) vs measured
cnn_dailymail_200eval-qwen-3max
litellm
coverage 0.3981$0.00082945 ms0—
cnn_qwen_m_n200_v1measuredcnn_dailymail_200eval-qwen-37plus
litellm
coverage 0.3031$0.00031534 ms1—
cnn_qwen_m_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-qwen-37plus
litellm
judge_won 0.0201$0.00031533 ms0—
cnn_qwen_m_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-qwen-37plus
litellm
judge_score 4.8761$0.00031532 ms0—
cnn_qwen_m_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-qwen-37plus
litellm
judge_score 4.6001$0.00031534 ms0—
cnn_qwen_m_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.303 → 0.381 (+0.078) vs measured
cnn_dailymail_200eval-qwen-37plus
litellm
coverage 0.3811$0.00031534 ms0—
cnn_qwen_s_n200_v1measuredcnn_dailymail_200eval-qwen-flash
litellm
coverage 0.3071$0.000125038 ms1—
cnn_qwen_s_n200_v1 · judged/judge-summ-pick-v1judged by judge-summ-pick-v1cnn_dailymail_200eval-qwen-flash
litellm
judge_won 0.0011$0.000125218 ms0—
cnn_qwen_s_n200_v1 · judged/judge-summ-v1judged by judge-summ-v1cnn_dailymail_200eval-qwen-flash
litellm
judge_score 4.6091$0.000124872 ms0—
cnn_qwen_s_n200_v1 · judged/judge-summ-v2judged by judge-summ-v2cnn_dailymail_200eval-qwen-flash
litellm
judge_score 4.4611$0.000125038 ms0—
cnn_qwen_s_n200_v1 · silverrescored against silver/cnn_dailymail_200
0.307 → 0.334 (+0.027) vs measured
cnn_dailymail_200eval-qwen-flash
litellm
coverage 0.3341$0.000125038 ms0—
cnn_bart_l_dm200_v1measuredcnn_dailymail_dm200facebook/bart-large-cnn
hf_local · local
coverage 0.4431—11260 ms1—
cnn_bart_l_xsum_dm200_v1measuredcnn_dailymail_dm200facebook/bart-large-xsum
hf_local · local
coverage 0.1521—5362 ms1—
cnn_bart_m_dm200_v1measuredcnn_dailymail_dm200sshleifer/distilbart-cnn-12-6
hf_local · local
coverage 0.4431—7292 ms1—
cnn_bart_s_dm200_v1measuredcnn_dailymail_dm200sshleifer/distilbart-cnn-6-6
hf_local · local
coverage 0.4051—4302 ms1—
cnn_lead3_dm200_v1measuredcnn_dailymail_dm200—
lead_k · local
coverage 0.3701—0 ms1—
db_anthropic_l_n200_v1measureddbpedia_280eval-opus-5
litellm
correct 0.9891$0.00143101 ms1—
db_anthropic_m_n200_v1measureddbpedia_280eval-claude-sonnet
litellm
correct 0.9821$0.00052258 ms1—
db_anthropic_s_n200_v1measureddbpedia_280eval-claude-haiku
litellm
correct 0.9791$0.00021083 ms1—
db_bart_mnli_n200_v1measureddbpedia_280facebook/bart-large-mnli
hf_zeroshot · local
correct 0.6291—5718 ms1—
db_bert_base_fy_n200_v1
baselinerecommended
measureddbpedia_280fabriceyhc/bert-base-uncased-dbpedia_14
hf_local · local
correct 0.9861—87 ms1—
db_bert_base_fy_n200_v1 · linux
retired in r2
run again (linux)
0.986 → 0.986 (+0.000) vs measured
dbpedia_280fabriceyhc/bert-base-uncased-dbpedia_14
hf_local · local
correct 0.9861—166 ms1—
db_constant_n200_v1measureddbpedia_280—
constant · local
correct 0.0711—0 ms1—
db_deepseek_l_n200_v1measureddbpedia_280eval-deepseek-pro
litellm
correct 0.9861$0.00011782 ms1—
db_deepseek_m_n200_v1measureddbpedia_280eval-deepseek-41flash
litellm
correct 0.9751$0.0000975 ms1—
db_deepseek_s_n200_v1measureddbpedia_280eval-deepseek-flash
litellm
correct 0.9431$0.00001548 ms1—
db_gemma_l_n200_v1measureddbpedia_280eval-gemma-31b
litellm
correct 0.9751$0.00001145 ms1—
db_gemma_m_n200_v1measureddbpedia_280eval-gemma-26b
litellm
correct 0.9861$0.0000856 ms1—
db_gemma_s_n200_v1measureddbpedia_280eval-gemma-3-27b
litellm
correct 0.9751$0.0000647 ms1—
db_glm_l_n200_v1measureddbpedia_280eval-glm-5
litellm
correct 0.9681$0.00011978 ms1—
db_glm_m_n200_v1measureddbpedia_280eval-glm-46
litellm
correct 0.9861$0.00011930 ms1—
db_glm_s_n200_v1measureddbpedia_280eval-glm-45-air
litellm
correct 0.9391$0.00001183 ms1—
db_keyword_n200_v1measureddbpedia_280—
keyword · local
correct 0.7071—0 ms1—
db_llama_l_n200_v1measureddbpedia_280eval-llama-maverick
litellm
correct 0.9751$0.0000660 ms1—
db_llama_m_n200_v1measureddbpedia_280eval-llama-70b
litellm
correct 0.9791$0.00001927 ms1—
db_llama_s_n200_v1measureddbpedia_280eval-llama-scout
litellm
correct 0.9571$0.00001374 ms1—
db_mistral_l_n200_v1measureddbpedia_280eval-mistral-large
litellm
correct 0.9641$0.000110565 ms1—
db_mistral_m_n200_v1measureddbpedia_280eval-mistral-medium
litellm
correct 0.9571$0.0001439 ms1—
db_mistral_s_n200_v1measureddbpedia_280eval-mistral-small
litellm
correct 0.9571$0.0000943 ms1—
db_openai_l_n200_v1measureddbpedia_280eval-gpt-6sol
litellm
correct 0.9751$0.00031266 ms1—
db_openai_m_n200_v1measureddbpedia_280eval-gpt-55
litellm
correct 0.9821$0.00091178 ms1—
db_openai_s_n200_v1measureddbpedia_280eval-gpt-mini
litellm
correct 0.9751$0.0001910 ms1—
db_qwen_l_n200_v1measureddbpedia_280eval-qwen-3max
litellm
correct 0.9681$0.00011660 ms1—
db_qwen_m_n200_v1
leader
measureddbpedia_280eval-qwen-37plus
litellm
correct 0.9931$0.0001853 ms1—
db_qwen_s_n200_v1measureddbpedia_280eval-qwen-flash
litellm
correct 0.9861$0.000012349 ms1—
fn_anthropic_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-opus-5
litellm
f1 0.6801$0.00252759 ms1—
fn_anthropic_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-claude-sonnet
litellm
f1 0.6661$0.00102692 ms1—
fn_anthropic_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-claude-haiku
litellm
f1 0.5841$0.00061338 ms1—
fn_capitalized_n200_v1rescored against gold/few_nerd_280few_nerd_280—
capitalized · local
f1 0.1911—0 ms1—
fn_deepseek_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-deepseek-pro
litellm
f1 0.5871$0.00033727 ms1—
fn_deepseek_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-deepseek-41flash
litellm
f1 0.6151$0.00011229 ms1—
fn_deepseek_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-deepseek-flash
litellm
f1 0.5761$0.00002277 ms1—
fn_gemma_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-gemma-31b
litellm
f1 0.6481$0.00002910 ms1—
fn_gemma_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-gemma-26b
litellm
f1 0.6731$0.00003050 ms1—
fn_gemma_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-gemma-3-27b
litellm
f1 0.5501$0.00005941 ms1—
fn_gliner_n200_v1rescored against gold/few_nerd_280few_nerd_280urchade/gliner_medium-v2.1
gliner · local
f1 0.4541—169 ms1—
fn_glm_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-glm-5
litellm
f1 0.5661$0.00043014 ms1—
fn_glm_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-glm-46
litellm
f1 0.5941$0.00023659 ms1—
fn_glm_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-glm-45-air
litellm
f1 0.5531$0.00012111 ms1—
fn_llama_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-llama-maverick
litellm
f1 0.5991$0.00014797 ms1—
fn_llama_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-llama-70b
litellm
f1 0.5741$0.00015402 ms1—
fn_llama_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-llama-scout
litellm
f1 0.5691$0.00002638 ms1—
fn_mistral_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-mistral-large
litellm
f1 0.5981$0.000110247 ms1—
fn_mistral_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-mistral-medium
litellm
f1 0.5911$0.0002794 ms1—
fn_mistral_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-mistral-small
litellm
f1 0.5241$0.00002064 ms1—
fn_nothing_n200_v1rescored against gold/few_nerd_280few_nerd_280—
nothing · local
f1 0.1251—0 ms1—
fn_openai_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-gpt-6sol
litellm
f1 0.6721$0.00061903 ms1—
fn_openai_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-gpt-55
litellm
f1 0.6861$0.00231598 ms1—
fn_openai_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-gpt-mini
litellm
f1 0.5951$0.00031206 ms1—
fn_qwen_l_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-qwen-3max
litellm
f1 0.6241$0.00031667 ms1—
fn_qwen_m_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-qwen-37plus
litellm
f1 0.6551$0.00012241 ms1—
fn_qwen_s_n200_v1rescored against gold/few_nerd_280few_nerd_280eval-qwen-flash
litellm
f1 0.6651$0.00012154 ms1—
fn_span_marker_n200_v1
baselineleaderrecommended
rescored against gold/few_nerd_280few_nerd_280guishe/span-marker-generic-ner-v1-fewnerd-fine-super
span_marker · local
f1 0.7671—758 ms1—
dedup_guard_v1measuredfleet_dedup_v1—
— · local
key_match 1.0001—0 ms1—
dedup_pinned_f07efa9_v1measuredfleet_dedup_v1—
— · local
—1—0 ms1—
triage_baseline_pro_noreason_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro @ hosted smoked
gateway
triage_cost 1.1541$0.00498352 ms1—
triage_baseline_pro_noreason_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
1.154 → 1.154 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro @ hosted smoked
gateway
triage_cost 1.1541$0.00498352 ms0—
triage_baseline_pro_noreason_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
1.154 → 1.154 (+0.000) vs measured
orrery_triage_v1_labelledeval-fleet-triage-pro @ hosted smoked
gateway
triage_cost 1.1541$0.00498352 ms0—
triage_baseline_pro_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
false_dismiss 0.1761$0.005022736 ms1—
triage_baseline_pro_v1 · repeats ×3run again (repeats)
0.176 → 0.167 (-0.010) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
false_dismiss 0.1673$0.005126285 ms3—
triage_baseline_pro_v1 · rescored 3d3a6afd, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7643$0.005126285 ms0—
triage_baseline_pro_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7851$0.005022736 ms0—
triage_baseline_pro_v1 · rescored 5f2c057c, repeats ×3rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7643$0.005126285 ms0—
triage_baseline_pro_v1 · rescored 5f2c057c
baseline
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7851$0.005022736 ms0—
triage_baseline_pro_v1 · rescored 77adc97b, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7643$0.005126285 ms0—
triage_baseline_pro_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7851$0.005022736 ms0—
triage_baseline_pro_v1 · rescored cd35dc1f, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7643$0.005126285 ms0—
triage_baseline_pro_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-fleet-triage-pro
gateway
triage_cost 0.7851$0.005022736 ms0—
triage_deepseek_41flash_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledeval-deepseek-41flash
gateway
false_dismiss 0.1761$0.00099833 ms1—
triage_deepseek_41flash_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-deepseek-41flash
gateway
triage_cost 0.7541$0.00099833 ms0—
triage_deepseek_41flash_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-deepseek-41flash
gateway
triage_cost 0.7541$0.00099833 ms0—
triage_deepseek_41flash_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-deepseek-41flash
gateway
triage_cost 0.7541$0.00099833 ms0—
triage_deepseek_41flash_v2measured
the goals read its re-score
orrery_triage_v1_labelledeval-deepseek-41flash @ hosted smoked
gateway
triage_cost 0.9081$0.00043282 ms1—
triage_deepseek_41flash_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.908 → 0.908 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-deepseek-41flash @ hosted smoked
gateway
triage_cost 0.9081$0.00043282 ms0—
triage_deepseek_41flash_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.908 → 0.908 (+0.000) vs measured
orrery_triage_v1_labelledeval-deepseek-41flash @ hosted smoked
gateway
triage_cost 0.9081$0.00043282 ms0—
triage_gemini_35_flash_lite_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-35-flash-lite
gateway
false_dismiss 0.0881$0.00143487 ms1—
triage_gemini_35_flash_lite_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-35-flash-lite
gateway
triage_cost 0.5381$0.00143487 ms0—
triage_gemini_35_flash_lite_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gemini-35-flash-lite
gateway
triage_cost 0.5381$0.00143487 ms0—
triage_gemini_35_flash_lite_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-35-flash-lite
gateway
triage_cost 0.5381$0.00143487 ms0—
triage_gemini_35_flash_lite_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-35-flash-lite
gateway
triage_cost 0.5381$0.00143487 ms0—
triage_gemini_4_argon_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-4-argon
gateway
false_dismiss 0.1181$0.013630118 ms1—
triage_gemini_4_argon_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-4-argon
gateway
triage_cost 0.6461$0.013630118 ms0—
triage_gemini_4_argon_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gemini-4-argon
gateway
triage_cost 0.6461$0.013630118 ms0—
triage_gemini_4_argon_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-4-argon
gateway
triage_cost 0.6461$0.013630118 ms0—
triage_gemini_4_argon_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemini-4-argon
gateway
triage_cost 0.6461$0.013630118 ms0—
triage_gpt_61_sol_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
false_dismiss 0.1181$0.010812241 ms1—
triage_gpt_61_sol_v1 · repeats ×3run again (repeats)
0.118 → 0.088 (-0.029) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
false_dismiss 0.0883$0.010912006 ms3—
triage_gpt_61_sol_v1 · rescored 3d3a6afd, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.5853$0.010912006 ms0—
triage_gpt_61_sol_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.6621$0.010812241 ms0—
triage_gpt_61_sol_v1 · rescored 5f2c057c, repeats ×3rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.5853$0.010912006 ms0—
triage_gpt_61_sol_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.6621$0.010812241 ms0—
triage_gpt_61_sol_v1 · rescored 77adc97b, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.5853$0.010912006 ms0—
triage_gpt_61_sol_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.6621$0.010812241 ms0—
triage_gpt_61_sol_v1 · rescored cd35dc1f, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.5853$0.010912006 ms0—
triage_gpt_61_sol_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-61-sol
gateway
triage_cost 0.6621$0.010812241 ms0—
triage_gpt_6_luna_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna
gateway
false_dismiss 0.0881$0.000712200 ms1—
triage_gpt_6_luna_v1 · repeats ×3run again (repeats)
0.088 → 0.118 (+0.029) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna
gateway
false_dismiss 0.1183$0.000711598 ms3—
triage_gpt_6_luna_v1 · rescored 5f2c057c, repeats ×3rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gpt-6-luna
gateway
triage_cost 0.6263$0.000711598 ms0—
triage_gpt_6_luna_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gpt-6-luna
gateway
triage_cost 0.5851$0.000712200 ms0—
triage_gpt_6_luna_v1 · rescored 77adc97b, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna
gateway
triage_cost 0.6263$0.000711598 ms0—
triage_gpt_6_luna_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna
gateway
triage_cost 0.5851$0.000712200 ms0—
triage_gpt_6_luna_v1 · rescored cd35dc1f, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna
gateway
triage_cost 0.6263$0.000711598 ms0—
triage_gpt_6_luna_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna
gateway
triage_cost 0.5851$0.000712200 ms0—
triage_gpt_6_luna_v2measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna @ hosted smoked
gateway
triage_cost 0.5541$0.00077973 ms1—
triage_gpt_6_luna_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.554 → 0.554 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-6-luna @ hosted smoked
gateway
triage_cost 0.5541$0.00077973 ms0—
triage_gpt_6_luna_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.554 → 0.554 (+0.000) vs measured
orrery_triage_v1_labelledeval-gpt-6-luna @ hosted smoked
gateway
triage_cost 0.5541$0.00077973 ms0—
triage_haiku_55_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
false_dismiss 0.1181$0.00107626 ms1—
triage_haiku_55_v1 · repeats ×3run again (repeats)
0.118 → 0.118 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
false_dismiss 0.1183$0.00107529 ms3—
triage_haiku_55_v1 · rescored 3d3a6afd, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.6103$0.00107529 ms0—
triage_haiku_55_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.5851$0.00107626 ms0—
triage_haiku_55_v1 · rescored 5f2c057c, repeats ×3rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.6103$0.00107529 ms0—
triage_haiku_55_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.5851$0.00107626 ms0—
triage_haiku_55_v1 · rescored 77adc97b, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.6103$0.00107529 ms0—
triage_haiku_55_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.5851$0.00107626 ms0—
triage_haiku_55_v1 · rescored cd35dc1f, repeats ×3rescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.6103$0.00107529 ms0—
triage_haiku_55_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-haiku-55
gateway
triage_cost 0.5851$0.00107626 ms0—
triage_hosted_gemma4_31b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledeval-gemma-31b
gateway
false_dismiss 0.1761$0.000910510 ms1—
triage_hosted_gemma4_31b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gemma-31b
gateway
triage_cost 0.7231$0.000910510 ms0—
triage_hosted_gemma4_31b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemma-31b
gateway
triage_cost 0.7231$0.000910510 ms0—
triage_hosted_gemma4_31b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gemma-31b
gateway
triage_cost 0.7231$0.000910510 ms0—
triage_hosted_gemma4_31b_v2measured
the goals read its re-score
orrery_triage_v1_labelledeval-gemma-31b @ hosted smoked
gateway
triage_cost 0.7231$0.000910550 ms1—
triage_hosted_gemma4_31b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-gemma-31b @ hosted smoked
gateway
triage_cost 0.7231$0.000910550 ms0—
triage_hosted_gemma4_31b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
orrery_triage_v1_labelledeval-gemma-31b @ hosted smoked
gateway
triage_cost 0.7231$0.000910550 ms0—
triage_hosted_gpt_oss_120b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-oss-120b
gateway
false_dismiss 0.1471$0.000817021 ms1—
triage_hosted_gpt_oss_120b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-gpt-oss-120b
gateway
triage_cost 0.6771$0.000817021 ms0—
triage_hosted_gpt_oss_120b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-oss-120b
gateway
triage_cost 0.6771$0.000817021 ms0—
triage_hosted_gpt_oss_120b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-oss-120b
gateway
triage_cost 0.6771$0.000817021 ms0—
triage_hosted_gpt_oss_120b_v2measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-oss-120b @ hosted smoked
gateway
triage_cost 0.6311$0.00066541 ms1—
triage_hosted_gpt_oss_120b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.631 → 0.631 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-gpt-oss-120b @ hosted smoked
gateway
triage_cost 0.6311$0.00066541 ms0—
triage_hosted_gpt_oss_120b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.631 → 0.631 (+0.000) vs measured
orrery_triage_v1_labelledeval-gpt-oss-120b @ hosted smoked
gateway
triage_cost 0.6311$0.00066541 ms0—
triage_hosted_mistral_medium_35_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-medium-35
gateway
false_dismiss 0.2351$0.01156103 ms1—
triage_hosted_mistral_medium_35_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-medium-35
gateway
triage_cost 0.8771$0.01156103 ms0—
triage_hosted_mistral_medium_35_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-mistral-medium-35
gateway
triage_cost 0.8771$0.01156103 ms0—
triage_hosted_mistral_medium_35_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-medium-35
gateway
triage_cost 0.8771$0.01156103 ms0—
triage_hosted_mistral_medium_35_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-medium-35
gateway
triage_cost 0.8771$0.01156103 ms0—
triage_hosted_qwen35_122b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledeval-qwen35-122b
gateway
false_dismiss 0.2941$0.003317662 ms1—
triage_hosted_qwen35_122b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-qwen35-122b
gateway
triage_cost 1.0771$0.003317662 ms0—
triage_hosted_qwen35_122b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-qwen35-122b
gateway
triage_cost 1.0771$0.003317662 ms0—
triage_hosted_qwen35_122b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-qwen35-122b
gateway
triage_cost 1.0771$0.003317662 ms0—
triage_hosted_qwen35_122b_v2measured
the goals read its re-score
orrery_triage_v1_labelledeval-qwen35-122b @ hosted smoked
gateway
triage_cost 0.8771$0.00176681 ms1—
triage_hosted_qwen35_122b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.877 → 0.877 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-qwen35-122b @ hosted smoked
gateway
triage_cost 0.8771$0.00176681 ms0—
triage_hosted_qwen35_122b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.877 → 0.877 (+0.000) vs measured
orrery_triage_v1_labelledeval-qwen35-122b @ hosted smoked
gateway
triage_cost 0.8771$0.00176681 ms0—
triage_hosted_qwen3_30b_a3b_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-qwen3-30b-a3b
gateway
triage_cost 0.7081$0.000717976 ms1—
triage_hosted_qwen3_30b_a3b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-qwen3-30b-a3b
gateway
triage_cost 0.7081$0.000717976 ms0—
triage_hosted_qwen3_30b_a3b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
orrery_triage_v1_labelledeval-qwen3-30b-a3b
gateway
triage_cost 0.7081$0.000717976 ms0—
triage_hosted_qwen3_30b_a3b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-qwen3-30b-a3b
gateway
triage_cost 0.7081$0.000717976 ms0—
triage_hosted_qwen3_30b_a3b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-qwen3-30b-a3b
gateway
triage_cost 0.7081$0.000717976 ms0—
triage_kimi_k31_pro_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledeval-kimi-k31-pro
gateway
false_dismiss 0.1181$0.019146089 ms1—
triage_kimi_k31_pro_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-kimi-k31-pro
gateway
triage_cost 0.6461$0.019146089 ms0—
triage_kimi_k31_pro_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-kimi-k31-pro
gateway
triage_cost 0.6461$0.019146089 ms0—
triage_kimi_k31_pro_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-kimi-k31-pro
gateway
triage_cost 0.6461$0.019146089 ms0—
triage_kimi_k31_pro_v2measured
the goals read its re-score
orrery_triage_v1_labelledeval-kimi-k31-pro @ hosted smoked
gateway
triage_cost 0.7231$0.012021533 ms1—
triage_kimi_k31_pro_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-kimi-k31-pro @ hosted smoked
gateway
triage_cost 0.7231$0.012021533 ms0—
triage_kimi_k31_pro_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
orrery_triage_v1_labelledeval-kimi-k31-pro @ hosted smoked
gateway
triage_cost 0.7231$0.012021533 ms0—
triage_local_dgx_gemma4_12b_noreason_v1measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—5697 ms1—
triage_local_dgx_gemma4_12b_noreason_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—5697 ms0—
triage_local_dgx_gemma4_12b_noreason_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—5697 ms0—
triage_local_dgx_gemma4_12b_noreason_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—5697 ms0—
triage_local_dgx_gemma4_12b_noreason_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—5697 ms0—
triage_local_dgx_gemma4_31b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6771—58863 ms1—
triage_local_dgx_gemma4_31b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6771—58863 ms0—
triage_local_dgx_gemma4_31b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6771—58863 ms0—
triage_local_dgx_gemma4_31b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6771—58863 ms0—
triage_local_dgx_gemma4_31b_v2measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b @ dgx smoked
gateway
triage_cost 0.7231—13030 ms1—
triage_local_dgx_gemma4_31b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b @ dgx smoked
gateway
triage_cost 0.7231—13030 ms0—
triage_local_dgx_gemma4_31b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:31b @ dgx smoked
gateway
triage_cost 0.7231—13030 ms0—
triage_local_dgx_gpt_oss_120b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:120b
gateway
triage_cost 0.6921—14766 ms1—
triage_local_dgx_gpt_oss_120b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
orrery_triage_v1_labelledgpt-oss:120b
gateway
triage_cost 0.6921—14766 ms0—
triage_local_dgx_gpt_oss_120b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:120b
gateway
triage_cost 0.6921—14766 ms0—
triage_local_dgx_gpt_oss_120b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:120b
gateway
triage_cost 0.6921—14766 ms0—
triage_local_dgx_gpt_oss_120b_v2measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:120b @ dgx smoke: check
gateway
triage_cost 0.6621—7391 ms1—
triage_local_dgx_gpt_oss_120b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.662 → 0.662 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:120b @ dgx smoke: check
gateway
triage_cost 0.6621—7391 ms0—
triage_local_dgx_gpt_oss_120b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.662 → 0.662 (+0.000) vs measured
orrery_triage_v1_labelledgpt-oss:120b @ dgx smoke: check
gateway
triage_cost 0.6621—7391 ms0—
triage_local_dgx_gpt_oss_20b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 0.4771—13456 ms1—
triage_local_dgx_gpt_oss_20b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.477 → 1.015 (+0.538) vs measured
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 1.0151—13456 ms0—
triage_local_dgx_gpt_oss_20b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.477 → 1.015 (+0.538) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 1.0151—13456 ms0—
triage_local_dgx_gpt_oss_20b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.477 → 0.477 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 0.4771—13456 ms0—
triage_local_dgx_gpt_oss_20b_v2measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b @ dgx smoke: check
gateway
triage_cost 0.7231—4681 ms1—
triage_local_dgx_gpt_oss_20b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b @ dgx smoke: check
gateway
triage_cost 0.7231—4681 ms0—
triage_local_dgx_gpt_oss_20b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
orrery_triage_v1_labelledgpt-oss:20b @ dgx smoke: check
gateway
triage_cost 0.7231—4681 ms0—
triage_local_dgx_llama31_8b_v1measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5766 ms1—
triage_local_dgx_llama31_8b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5766 ms0—
triage_local_dgx_llama31_8b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5766 ms0—
triage_local_dgx_llama31_8b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5766 ms0—
triage_local_dgx_llama31_8b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5766 ms0—
triage_local_dgx_llama4_scout_v1measured
the goals read its re-score
orrery_triage_v1_labelledllama4:scout
gateway
triage_cost 1.2621—13354 ms1—
triage_local_dgx_llama4_scout_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
1.262 → 1.262 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama4:scout
gateway
triage_cost 1.2621—13354 ms0—
triage_local_dgx_llama4_scout_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
1.262 → 1.262 (+0.000) vs measured
orrery_triage_v1_labelledllama4:scout
gateway
triage_cost 1.2621—13354 ms0—
triage_local_dgx_llama4_scout_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
1.262 → 1.262 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama4:scout
gateway
triage_cost 1.2621—13354 ms0—
triage_local_dgx_llama4_scout_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
1.262 → 1.262 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama4:scout
gateway
triage_cost 1.2621—13354 ms0—
triage_local_dgx_ministral_3_14b_v1measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8311—20452 ms1—
triage_local_dgx_ministral_3_14b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8311—20452 ms0—
triage_local_dgx_ministral_3_14b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8311—20452 ms0—
triage_local_dgx_ministral_3_14b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8311—20452 ms0—
triage_local_dgx_ministral_3_14b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8311—20452 ms0—
triage_local_dgx_mistral_medium_35_128b_v1measured
the goals read its re-score
orrery_triage_v1_labelledmistral-medium-3.5:128b
gateway
triage_cost 1.0001—117538 ms1—
triage_local_dgx_mistral_medium_35_128b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
1.000 → 1.000 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledmistral-medium-3.5:128b
gateway
triage_cost 1.0001—117538 ms0—
triage_local_dgx_mistral_medium_35_128b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
1.000 → 1.000 (+0.000) vs measured
orrery_triage_v1_labelledmistral-medium-3.5:128b
gateway
triage_cost 1.0001—117538 ms0—
triage_local_dgx_mistral_medium_35_128b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
1.000 → 1.000 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledmistral-medium-3.5:128b
gateway
triage_cost 1.0001—117538 ms0—
triage_local_dgx_mistral_medium_35_128b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
1.000 → 1.000 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledmistral-medium-3.5:128b
gateway
triage_cost 1.0001—117538 ms0—
triage_local_dgx_nemotron_35_bolt_e4b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.5231—15259 ms1—
triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.523 → 0.523 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.5231—15259 ms0—
triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.523 → 0.523 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.5231—15259 ms0—
triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.523 → 0.523 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.5231—15259 ms0—
triage_local_dgx_nemotron_35_bolt_e4b_v2measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ dgx smoke: check
gateway
triage_cost 0.6461—3881 ms1—
triage_local_dgx_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ dgx smoke: check
gateway
triage_cost 0.6461—3881 ms0—
triage_local_dgx_nemotron_35_bolt_e4b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ dgx smoke: check
gateway
triage_cost 0.6461—3881 ms0—
triage_local_dgx_nemotron_35_lightning_30b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.6461—39979 ms1—
triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.6461—39979 ms0—
triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.6461—39979 ms0—
triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.6461—39979 ms0—
triage_local_dgx_nemotron_35_lightning_30b_v2measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b @ dgx smoked
gateway
triage_cost 0.6771—6786 ms1—
triage_local_dgx_nemotron_35_lightning_30b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b @ dgx smoked
gateway
triage_cost 0.6771—6786 ms0—
triage_local_dgx_nemotron_35_lightning_30b_v2 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-lightning:30b @ dgx smoked
gateway
triage_cost 0.6771—6786 ms0—
triage_local_dgx_nemotron_35_super_120b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-super:120b
gateway
triage_cost 0.6771—41019 ms1—
triage_local_dgx_nemotron_35_super_120b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-super:120b
gateway
triage_cost 0.6771—41019 ms0—
triage_local_dgx_nemotron_35_super_120b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-super:120b
gateway
triage_cost 0.6771—41019 ms0—
triage_local_dgx_nemotron_35_super_120b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-super:120b
gateway
triage_cost 0.6771—41019 ms0—
triage_local_dgx_nemotron_35_super_120b_v2measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-super:120b @ dgx smoked
gateway
triage_cost 0.6771—15898 ms1—
triage_local_dgx_nemotron_35_super_120b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-super:120b @ dgx smoked
gateway
triage_cost 0.6771—15898 ms0—
triage_local_dgx_nemotron_35_super_120b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-super:120b @ dgx smoked
gateway
triage_cost 0.6771—15898 ms0—
triage_local_dgx_phi4_14b_v1measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.6621—16973 ms1—
triage_local_dgx_phi4_14b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.662 → 0.662 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.6621—16973 ms0—
triage_local_dgx_phi4_14b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.662 → 0.662 (+0.000) vs measured
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.6621—16973 ms0—
triage_local_dgx_phi4_14b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.662 → 0.662 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.6621—16973 ms0—
triage_local_dgx_phi4_14b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.662 → 0.662 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.6621—16973 ms0—
triage_local_dgx_qwen35_122b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.5:122b
gateway
triage_cost 1.1381—38007 ms1—
triage_local_dgx_qwen35_122b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
1.138 → 1.138 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.5:122b
gateway
triage_cost 1.1381—38007 ms0—
triage_local_dgx_qwen35_122b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
1.138 → 1.138 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.5:122b
gateway
triage_cost 1.1381—38007 ms0—
triage_local_dgx_qwen35_122b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
1.138 → 1.138 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.5:122b
gateway
triage_cost 1.1381—38007 ms0—
triage_local_dgx_qwen35_122b_v2measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.5:122b @ dgx smoked
gateway
triage_cost 1.3541—14511 ms1—
triage_local_dgx_qwen35_122b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
1.354 → 1.354 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.5:122b @ dgx smoked
gateway
triage_cost 1.3541—14511 ms0—
triage_local_dgx_qwen35_122b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
1.354 → 1.354 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.5:122b @ dgx smoked
gateway
triage_cost 1.3541—14511 ms0—
triage_local_dgx_qwen38_27b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.8311—37826 ms1—
triage_local_dgx_qwen38_27b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.8311—37826 ms0—
triage_local_dgx_qwen38_27b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.8311—37826 ms0—
triage_local_dgx_qwen38_27b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.8311—37826 ms0—
triage_local_dgx_qwen38_27b_v2measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b @ dgx smoked
gateway
triage_cost 1.1381—16559 ms1—
triage_local_dgx_qwen38_27b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
1.138 → 1.138 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b @ dgx smoked
gateway
triage_cost 1.1381—16559 ms0—
triage_local_dgx_qwen38_27b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
1.138 → 1.138 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.8:27b @ dgx smoked
gateway
triage_cost 1.1381—16559 ms0—
triage_local_dgx_qwen3_30b_a3b_v1measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8921—5859 ms1—
triage_local_dgx_qwen3_30b_a3b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.892 → 0.892 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8921—5859 ms0—
triage_local_dgx_qwen3_30b_a3b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.892 → 0.892 (+0.000) vs measured
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8921—5859 ms0—
triage_local_dgx_qwen3_30b_a3b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.892 → 0.892 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8921—5859 ms0—
triage_local_dgx_qwen3_30b_a3b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.892 → 0.892 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8921—5859 ms0—
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1measured
the goals read its re-score
orrery_triage_v1_labelledNVFP4/Qwen3-30B-A3B-Instruct-2507-FP4
gateway
triage_cost 0.4001—7824 ms1—
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.400 → 0.400 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledNVFP4/Qwen3-30B-A3B-Instruct-2507-FP4
gateway
triage_cost 0.4001—7824 ms0—
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 5f2c057c
leader
rescored against gold/orrery_triage_v1_labelled
0.400 → 0.400 (+0.000) vs measured
orrery_triage_v1_labelledNVFP4/Qwen3-30B-A3B-Instruct-2507-FP4
gateway
triage_cost 0.4001—7824 ms0—
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.400 → 0.400 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledNVFP4/Qwen3-30B-A3B-Instruct-2507-FP4
gateway
triage_cost 0.4001—7824 ms0—
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.400 → 0.400 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledNVFP4/Qwen3-30B-A3B-Instruct-2507-FP4
gateway
triage_cost 0.4001—7824 ms0—
triage_local_mac_gemma4_12b_noreason_v1measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—9174 ms1—
triage_local_mac_gemma4_12b_noreason_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—9174 ms0—
triage_local_mac_gemma4_12b_noreason_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—9174 ms0—
triage_local_mac_gemma4_12b_noreason_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—9174 ms0—
triage_local_mac_gemma4_12b_noreason_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:12b
gateway
triage_cost 0.5851—9174 ms0—
triage_local_mac_gemma4_31b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b-it-q4_K_M
gateway
triage_cost 0.6771—100889 ms1—
triage_local_mac_gemma4_31b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:31b-it-q4_K_M
gateway
triage_cost 0.6771—100889 ms0—
triage_local_mac_gemma4_31b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b-it-q4_K_M
gateway
triage_cost 0.6771—100889 ms0—
triage_local_mac_gemma4_31b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b-it-q4_K_M
gateway
triage_cost 0.6771—100889 ms0—
triage_local_mac_gemma4_31b_v2measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b-it-q4_K_M @ mac smoked
gateway
triage_cost 0.8001—25063 ms1—
triage_local_mac_gemma4_31b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.800 → 0.800 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b-it-q4_K_M @ mac smoked
gateway
triage_cost 0.8001—25063 ms0—
triage_local_mac_gemma4_31b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.800 → 0.800 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:31b-it-q4_K_M @ mac smoked
gateway
triage_cost 0.8001—25063 ms0—
triage_local_mac_gpt_oss_20b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 0.5541—19953 ms1—
triage_local_mac_gpt_oss_20b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.554 → 1.092 (+0.538) vs measured
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 1.0921—19953 ms0—
triage_local_mac_gpt_oss_20b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.554 → 1.092 (+0.538) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 1.0921—19953 ms0—
triage_local_mac_gpt_oss_20b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.554 → 0.554 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 0.5541—19953 ms0—
triage_local_mac_gpt_oss_20b_v2measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b @ mac smoke: check
gateway
triage_cost 0.7081—5802 ms1—
triage_local_mac_gpt_oss_20b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b @ mac smoke: check
gateway
triage_cost 0.7081—5802 ms0—
triage_local_mac_gpt_oss_20b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
orrery_triage_v1_labelledgpt-oss:20b @ mac smoke: check
gateway
triage_cost 0.7081—5802 ms0—
triage_local_mac_llama31_8b_v1measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—7780 ms1—
triage_local_mac_llama31_8b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—7780 ms0—
triage_local_mac_llama31_8b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—7780 ms0—
triage_local_mac_llama31_8b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—7780 ms0—
triage_local_mac_llama31_8b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.415 → 0.415 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—7780 ms0—
triage_local_mac_ministral_3_14b_v1measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8311—27378 ms1—
triage_local_mac_ministral_3_14b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.831 → 0.908 (+0.077) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.9081—27378 ms0—
triage_local_mac_ministral_3_14b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.831 → 0.908 (+0.077) vs measured
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.9081—27378 ms0—
triage_local_mac_ministral_3_14b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.831 → 0.908 (+0.077) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.9081—27378 ms0—
triage_local_mac_ministral_3_14b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8311—27378 ms0—
triage_local_mac_nemotron_35_bolt_e4b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.7541—23126 ms1—
triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.754 → 0.754 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.7541—23126 ms0—
triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.754 → 0.754 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.7541—23126 ms0—
triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.754 → 0.754 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.7541—23126 ms0—
triage_local_mac_nemotron_35_bolt_e4b_v2measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ mac smoked
gateway
triage_cost 0.5851—5883 ms1—
triage_local_mac_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ mac smoked
gateway
triage_cost 0.5851—5883 ms0—
triage_local_mac_nemotron_35_bolt_e4b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.585 → 0.585 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ mac smoked
gateway
triage_cost 0.5851—5883 ms0—
triage_local_mac_nemotron_35_lightning_30b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b-a3b-q4_K_M
gateway
triage_cost 0.6461—66004 ms1—
triage_local_mac_nemotron_35_lightning_30b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-lightning:30b-a3b-q4_K_M
gateway
triage_cost 0.6461—66004 ms0—
triage_local_mac_nemotron_35_lightning_30b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b-a3b-q4_K_M
gateway
triage_cost 0.6461—66004 ms0—
triage_local_mac_nemotron_35_lightning_30b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.646 → 0.646 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b-a3b-q4_K_M
gateway
triage_cost 0.6461—66004 ms0—
triage_local_mac_nemotron_35_lightning_30b_v2measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b-a3b-q4_K_M @ mac smoked
gateway
triage_cost 0.6921—11065 ms1—
triage_local_mac_nemotron_35_lightning_30b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b-a3b-q4_K_M @ mac smoked
gateway
triage_cost 0.6921—11065 ms0—
triage_local_mac_nemotron_35_lightning_30b_v2 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-lightning:30b-a3b-q4_K_M @ mac smoked
gateway
triage_cost 0.6921—11065 ms0—
triage_local_mac_phi4_14b_v1measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7541—21773 ms1—
triage_local_mac_phi4_14b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.754 → 0.754 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7541—21773 ms0—
triage_local_mac_phi4_14b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.754 → 0.754 (+0.000) vs measured
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7541—21773 ms0—
triage_local_mac_phi4_14b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.754 → 0.754 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7541—21773 ms0—
triage_local_mac_phi4_14b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.754 → 0.754 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7541—21773 ms0—
triage_local_mac_qwen38_27b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 1.0461—69674 ms1—
triage_local_mac_qwen38_27b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
1.046 → 1.046 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 1.0461—69674 ms0—
triage_local_mac_qwen38_27b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
1.046 → 1.046 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 1.0461—69674 ms0—
triage_local_mac_qwen38_27b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
1.046 → 1.046 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 1.0461—69674 ms0—
triage_local_mac_qwen38_27b_v2measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b @ mac smoked
gateway
triage_cost 1.1381—33395 ms1—
triage_local_mac_qwen38_27b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
1.138 → 1.138 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b @ mac smoked
gateway
triage_cost 1.1381—33395 ms0—
triage_local_mac_qwen38_27b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
1.138 → 1.138 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.8:27b @ mac smoked
gateway
triage_cost 1.1381—33395 ms0—
triage_local_mac_qwen3_30b_a3b_v1measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8311—8996 ms1—
triage_local_mac_qwen3_30b_a3b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8311—8996 ms0—
triage_local_mac_qwen3_30b_a3b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8311—8996 ms0—
triage_local_mac_qwen3_30b_a3b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8311—8996 ms0—
triage_local_mac_qwen3_30b_a3b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.831 → 0.831 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.8311—8996 ms0—
triage_local_x3_gemma4_31b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6921—72058 ms1—
triage_local_x3_gemma4_31b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6921—72058 ms0—
triage_local_x3_gemma4_31b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6921—72058 ms0—
triage_local_x3_gemma4_31b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.692 → 0.692 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b
gateway
triage_cost 0.6921—72058 ms0—
triage_local_x3_gemma4_31b_v2measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b @ x3 smoke: check
gateway
triage_cost 0.7081—14320 ms1—
triage_local_x3_gemma4_31b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgemma4:31b @ x3 smoke: check
gateway
triage_cost 0.7081—14320 ms0—
triage_local_x3_gemma4_31b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.708 → 0.708 (+0.000) vs measured
orrery_triage_v1_labelledgemma4:31b @ x3 smoke: check
gateway
triage_cost 0.7081—14320 ms0—
triage_local_x3_gpt_oss_20b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 0.4001—17235 ms1—
triage_local_x3_gpt_oss_20b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.400 → 1.092 (+0.692) vs measured
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 1.0921—17235 ms0—
triage_local_x3_gpt_oss_20b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.400 → 1.092 (+0.692) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 1.0921—17235 ms0—
triage_local_x3_gpt_oss_20b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.400 → 0.400 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b
gateway
triage_cost 0.4001—17235 ms0—
triage_local_x3_gpt_oss_20b_v2measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b @ x3 smoke: check
gateway
triage_cost 0.8151—5864 ms1—
triage_local_x3_gpt_oss_20b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.815 → 0.815 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledgpt-oss:20b @ x3 smoke: check
gateway
triage_cost 0.8151—5864 ms0—
triage_local_x3_gpt_oss_20b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.815 → 0.815 (+0.000) vs measured
orrery_triage_v1_labelledgpt-oss:20b @ x3 smoke: check
gateway
triage_cost 0.8151—5864 ms0—
triage_local_x3_llama31_8b_v1measured
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
false_dismiss 0.0001—5837 ms1—
triage_local_x3_llama31_8b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5837 ms0—
triage_local_x3_llama31_8b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5837 ms0—
triage_local_x3_llama31_8b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5837 ms0—
triage_local_x3_llama31_8b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledllama3.1:8b
gateway
triage_cost 0.4151—5837 ms0—
triage_local_x3_ministral_3_14b_v1measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8151—22057 ms1—
triage_local_x3_ministral_3_14b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.815 → 0.815 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8151—22057 ms0—
triage_local_x3_ministral_3_14b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.815 → 0.815 (+0.000) vs measured
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8151—22057 ms0—
triage_local_x3_ministral_3_14b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.815 → 0.815 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8151—22057 ms0—
triage_local_x3_ministral_3_14b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.815 → 0.815 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledministral-3:14b
gateway
triage_cost 0.8151—22057 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—22252 ms1—
triage_local_x3_nemotron_35_bolt_e4b_v1 · repeatsrun again (repeats)
0.600 → 0.600 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—21266 ms1—
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c, repeatsrescored against gold/orrery_triage_v1_labelled
0.600 → 0.600 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—21266 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.600 → 0.600 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—22252 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 77adc97b, repeatsrescored against gold/orrery_triage_v1_labelled
0.600 → 0.600 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—21266 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.600 → 0.600 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—22252 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f, repeatsrescored against gold/orrery_triage_v1_labelled
0.600 → 0.600 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—21266 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.600 → 0.600 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M
gateway
triage_cost 0.6001—22252 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v2measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ x3 smoke: check
gateway
triage_cost 0.5541—5646 ms1—
triage_local_x3_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.554 → 0.554 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ x3 smoke: check
gateway
triage_cost 0.5541—5646 ms0—
triage_local_x3_nemotron_35_bolt_e4b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.554 → 0.554 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-bolt:e4b-q4_K_M @ x3 smoke: check
gateway
triage_cost 0.5541—5646 ms0—
triage_local_x3_nemotron_35_lightning_30b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.5231—39458 ms1—
triage_local_x3_nemotron_35_lightning_30b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.523 → 0.523 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.5231—39458 ms0—
triage_local_x3_nemotron_35_lightning_30b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.523 → 0.523 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.5231—39458 ms0—
triage_local_x3_nemotron_35_lightning_30b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.523 → 0.523 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b
gateway
triage_cost 0.5231—39458 ms0—
triage_local_x3_nemotron_35_lightning_30b_v2measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b @ x3 smoke: check
gateway
triage_cost 0.6771—7797 ms1—
triage_local_x3_nemotron_35_lightning_30b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labellednemotron-3.5-lightning:30b @ x3 smoke: check
gateway
triage_cost 0.6771—7797 ms0—
triage_local_x3_nemotron_35_lightning_30b_v2 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.677 → 0.677 (+0.000) vs measured
orrery_triage_v1_labellednemotron-3.5-lightning:30b @ x3 smoke: check
gateway
triage_cost 0.6771—7797 ms0—
triage_local_x3_phi4_14b_v1measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7691—16911 ms1—
triage_local_x3_phi4_14b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.769 → 0.769 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7691—16911 ms0—
triage_local_x3_phi4_14b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.769 → 0.769 (+0.000) vs measured
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7691—16911 ms0—
triage_local_x3_phi4_14b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.769 → 0.769 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7691—16911 ms0—
triage_local_x3_phi4_14b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.769 → 0.769 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledphi4:14b
gateway
triage_cost 0.7691—16911 ms0—
triage_local_x3_qwen38_27b_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.9081—46161 ms1—
triage_local_x3_qwen38_27b_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.908 → 0.908 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.9081—46161 ms0—
triage_local_x3_qwen38_27b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.908 → 0.908 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.9081—46161 ms0—
triage_local_x3_qwen38_27b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.908 → 0.908 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b
gateway
triage_cost 0.9081—46161 ms0—
triage_local_x3_qwen38_27b_v2measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b @ x3 smoke: check
gateway
triage_cost 1.2921—20630 ms1—
triage_local_x3_qwen38_27b_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
1.292 → 1.292 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3.8:27b @ x3 smoke: check
gateway
triage_cost 1.2921—20630 ms0—
triage_local_x3_qwen38_27b_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
1.292 → 1.292 (+0.000) vs measured
orrery_triage_v1_labelledqwen3.8:27b @ x3 smoke: check
gateway
triage_cost 1.2921—20630 ms0—
triage_local_x3_qwen3_30b_a3b_v1measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.7231—6932 ms1—
triage_local_x3_qwen3_30b_a3b_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.7231—6932 ms0—
triage_local_x3_qwen3_30b_a3b_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.7231—6932 ms0—
triage_local_x3_qwen3_30b_a3b_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.7231—6932 ms0—
triage_local_x3_qwen3_30b_a3b_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.723 → 0.723 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledqwen3:30b-a3b-instruct-2507-q4_K_M
gateway
triage_cost 0.7231—6932 ms0—
triage_mistral_large_4_v1
retired in r3
measured
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-large-4
gateway
false_dismiss 0.1761$0.006917290 ms1—
triage_mistral_large_4_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-mistral-large-4
gateway
triage_cost 0.8001$0.006917290 ms0—
triage_mistral_large_4_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-large-4
gateway
triage_cost 0.8001$0.006917290 ms0—
triage_mistral_large_4_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-large-4
gateway
triage_cost 0.8001$0.006917290 ms0—
triage_mistral_large_4_v2measured
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-large-4 @ hosted smoked
gateway
triage_cost 0.8001$0.006920489 ms1—
triage_mistral_large_4_v2 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.800 → 0.800 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelledeval-mistral-large-4 @ hosted smoked
gateway
triage_cost 0.8001$0.006920489 ms0—
triage_mistral_large_4_v2 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.800 → 0.800 (+0.000) vs measured
orrery_triage_v1_labelledeval-mistral-large-4 @ hosted smoked
gateway
triage_cost 0.8001$0.006920489 ms0—
triage_sonnet_55_v1measured
the goals read its re-score
orrery_triage_v1_labelledeval-claude-sonnet-55
gateway
false_dismiss 0.1471$0.022712035 ms1—
triage_sonnet_55_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-sonnet-55
gateway
triage_cost 0.7081$0.022712035 ms0—
triage_sonnet_55_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelledorrery_triage_v1_labelledeval-claude-sonnet-55
gateway
triage_cost 0.7081$0.022712035 ms0—
triage_sonnet_55_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-sonnet-55
gateway
triage_cost 0.7081$0.022712035 ms0—
triage_sonnet_55_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
the goals read its re-score
orrery_triage_v1_labelledeval-claude-sonnet-55
gateway
triage_cost 0.7081$0.022712035 ms0—
triage_stub_dismiss_v1measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.8621—0 ms1—
triage_stub_dismiss_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.862 → 0.862 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.8621—0 ms0—
triage_stub_dismiss_v1 · rescored 5f2c057crescored against gold/orrery_triage_v1_labelled
0.862 → 0.862 (+0.000) vs measured
orrery_triage_v1_labelled—
stub · local
triage_cost 0.8621—0 ms0—
triage_stub_dismiss_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.862 → 0.862 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.8621—0 ms0—
triage_stub_dismiss_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.862 → 0.862 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.8621—0 ms0—
triage_stub_escalate_v1measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.4771—0 ms1—
triage_stub_escalate_v1 · rescored 3d3a6afdrescored against gold/orrery_triage_v1_labelled
0.477 → 0.477 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.4771—0 ms0—
triage_stub_escalate_v1 · rescored 5f2c057c
not eligible
rescored against gold/orrery_triage_v1_labelled
0.477 → 0.477 (+0.000) vs measured
orrery_triage_v1_labelled—
stub · local
triage_cost 0.4771—0 ms0—
triage_stub_escalate_v1 · rescored 77adc97brescored against gold/orrery_triage_v1_labelled
0.477 → 0.477 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.4771—0 ms0—
triage_stub_escalate_v1 · rescored cd35dc1frescored against gold/orrery_triage_v1_labelled
0.477 → 0.477 (+0.000) vs measured
the goals read its re-score
orrery_triage_v1_labelled—
stub · local
triage_cost 0.4771—0 ms0—
sf_bge_small_n200_v1re-parsedscifact_200BAAI/bge-small-en-v1.5
dense · local
ndcg_10 0.7101—29 ms1—
sf_bm25_n200_v1re-parsedscifact_200—
bm25 · local
ndcg_10 0.6451—22 ms1—
sf_deepseek_m_n200_v1re-parsedscifact_200eval-deepseek-41flash
rerank
ndcg_10 0.7401$0.00121145 ms1—
sf_deepseek_s_n200_v1re-parsedscifact_200eval-deepseek-flash
rerank
ndcg_10 0.7381$0.00033453 ms1—
sf_e5_base_n200_v1
baselinerecommended
re-parsedscifact_200intfloat/e5-base-v2
dense · local
ndcg_10 0.7191—59 ms1—
sf_first_k_n200_v1re-parsedscifact_200—
first_k · local
ndcg_10 0.0001—0 ms1—
sf_gemma_l_n200_v1re-parsedscifact_200eval-gemma-31b
rerank
ndcg_10 0.7361$0.00099701 ms1—
sf_gemma_m_n200_v1re-parsedscifact_200eval-gemma-26b
rerank
ndcg_10 0.7361$0.00045664 ms1—
sf_gemma_s_n200_v1re-parsedscifact_200eval-gemma-3-27b
rerank
ndcg_10 0.7021$0.00057071 ms1—
sf_glm_m_n200_v1re-parsedscifact_200eval-glm-46
rerank
ndcg_10 0.7421$0.00236920 ms1—
sf_glm_s_bm25first_n200_v1 · pair
retired in r2
re-parsedscifact_200eval-glm-45-air
rerank
ndcg_10 0.7421$0.00073024 ms1—
sf_glm_s_e5first_n200_v1 · pair
retired in r2
re-parsedscifact_200eval-glm-45-air
rerank
ndcg_10 0.7891$0.00073127 ms1—
sf_glm_s_n200_v1
leader
re-parsedscifact_200eval-glm-45-air
rerank
ndcg_10 0.7441$0.00073120 ms1—
sf_llama_l_n200_v1re-parsedscifact_200eval-llama-maverick
rerank
ndcg_10 0.7281$0.00094554 ms1—
sf_llama_l_n200_v1 · repeats
retired in r3
re-parsedscifact_200eval-llama-maverick
rerank
ndcg_10 0.7151$0.000811386 ms1—
sf_llama_m_n200_v1re-parsedscifact_200eval-llama-70b
rerank
ndcg_10 0.7091$0.00074064 ms1—
sf_llama_s_n200_v1re-parsedscifact_200eval-llama-scout
rerank
ndcg_10 0.7171$0.000717482 ms1—
sf_llama_s_n200_v1 · repeats
retired in r3
re-parsedscifact_200eval-llama-scout
rerank
ndcg_10 0.7131$0.00054874 ms1—
sf_minilm_n200_v1re-parsedscifact_200sentence-transformers/all-MiniLM-L6-v2
dense · local
ndcg_10 0.6531—17 ms1—
sf_mistral_s_n200_v1re-parsedscifact_200eval-mistral-small
rerank
ndcg_10 0.7071$0.00047224 ms1—
sf_mpnet_n200_v1re-parsedscifact_200sentence-transformers/all-mpnet-base-v2
dense · local
ndcg_10 0.6641—65 ms1—
sf_qwen_s_bm25first_n200_v1 · pair
retired in r2
re-parsedscifact_200eval-qwen-flash
rerank
ndcg_10 0.7141$0.00086196 ms1—
sf_qwen_s_e5first_n200_v1 · pair
retired in r2
re-parsedscifact_200eval-qwen-flash
rerank
ndcg_10 0.7711$0.00086145 ms1—
sf_qwen_s_n200_v1re-parsedscifact_200eval-qwen-flash
rerank
ndcg_10 0.7181$0.00085651 ms1—
sf_qwen_s_n200_v1 · repeats
retired in r3
re-parsedscifact_200eval-qwen-flash
rerank
ndcg_10 0.7191$0.00085976 ms1—
sf_random_n200_v1re-parsedscifact_200—
random · local
ndcg_10 0.0011—2 ms1—
demo_v1measuredsmoke_v1—
echo · local
overlap_f1 0.4161—0 ms1—
se_anthropic_l_n100_v1measuredsummeval_100eval-opus-5
litellm
tau 0.5281$0.02013427 ms1—
se_anthropic_m_n100_v1measuredsummeval_100eval-claude-sonnet
litellm
tau 0.4801$0.00803008 ms1—
se_anthropic_s_n100_v1measuredsummeval_100eval-claude-haiku
litellm
tau 0.4841$0.00322433 ms1—
se_bartscore_n100_v1measuredsummeval_100facebook/bart-large-cnn
bartscore · local
tau 0.2561—15662 ms1—
se_coverage_n100_v1measuredsummeval_100—
coverage · local
tau 0.1011—892 ms1—
se_deepseek_l_n100_v1measuredsummeval_100eval-deepseek-pro
litellm
tau 0.4691$0.00035535 ms1—
se_deepseek_m_n100_v1measuredsummeval_100eval-deepseek-41flash
litellm
tau 0.4861$0.00081120 ms1—
se_deepseek_s_n100_v1measuredsummeval_100eval-deepseek-flash
litellm
tau 0.4591$0.00033985 ms1—
se_gemma_l_n100_pairwise_v1measuredsummeval_100eval-gemma-31b
litellm
pair_agreement 0.8251$0.00086981 ms1—
se_gemma_l_n100_pick_v1measuredsummeval_100eval-gemma-31b
litellm
top1 0.3641$0.00021813 ms1—
se_gemma_l_n100_v1
baselinerecommended
measuredsummeval_100eval-gemma-31b
litellm
tau 0.5381$0.00036670 ms1—
se_gemma_l_pubtext_n100_v1 · pair
retired in r2
run again (pair)summeval_100eval-gemma-31b
litellm
tau 0.5511$0.00102614 ms1—
se_gemma_l_pubtextcw_n100_v1 · pair
retired in r2
run again (pair)summeval_100eval-gemma-31b
litellm
tau 0.5401$0.00033268 ms1—
se_gemma_m_n100_v1measuredsummeval_100eval-gemma-26b
litellm
tau 0.4771$0.00034308 ms1—
se_gemma_s_n100_v1measuredsummeval_100eval-gemma-3-27b
litellm
tau 0.4201$0.000311572 ms1—
se_glm_l_n100_v1measuredsummeval_100eval-glm-5
litellm
tau 0.4921$0.00253590 ms1—
se_glm_m_n100_v1measuredsummeval_100eval-glm-46
litellm
tau 0.5171$0.00113609 ms1—
se_glm_s_n100_v1measuredsummeval_100eval-glm-45-air
litellm
tau 0.4461$0.00055617 ms1—
se_length_n100_v1measuredsummeval_100—
length · local
tau 0.0391—0 ms1—
se_llama_l_n100_v1measuredsummeval_100eval-llama-maverick
litellm
tau 0.4541$0.000612346 ms1—
se_llama_m_n100_v1measuredsummeval_100eval-llama-70b
litellm
tau 0.4061$0.00054834 ms1—
se_llama_s_n100_v1measuredsummeval_100eval-llama-scout
litellm
tau 0.3531$0.00037141 ms1—
se_mistral_l_n100_v1measuredsummeval_100eval-mistral-large
litellm
tau 0.4551$0.001310885 ms1—
se_mistral_m_n100_v1measuredsummeval_100eval-mistral-medium
litellm
tau 0.4681$0.00121946 ms1—
se_mistral_s_n100_v1measuredsummeval_100eval-mistral-small
litellm
tau 0.4271$0.00039185 ms1—
se_openai_l_n100_v1
leader
measuredsummeval_100eval-gpt-6sol
litellm
tau 0.5401$0.00632228 ms1—
se_openai_m_n100_v1measuredsummeval_100eval-gpt-55
litellm
tau 0.5281$0.01422690 ms1—
se_openai_s_n100_v1measuredsummeval_100eval-gpt-mini
litellm
tau 0.4111$0.00211551 ms1—
se_qwen37max_n100_v1measuredsummeval_100eval-qwen-37max
litellm
tau 0.5201$0.00444514 ms1—
se_qwen_m_n100_v1measuredsummeval_100eval-qwen-37plus
litellm
tau 0.5071$0.00116770 ms1—
se_qwen_s_n100_pairwise_v1measuredsummeval_100eval-qwen-flash
litellm
pair_agreement 0.8121$0.000520081 ms1—
se_qwen_s_n100_pick_v1measuredsummeval_100eval-qwen-flash
litellm
top1 0.2631$0.00031492 ms1—
se_qwen_s_n100_v1measuredsummeval_100eval-qwen-flash
litellm
tau 0.5191$0.00043954 ms1—
se_random_n100_pairwise_v1measuredsummeval_100—
random · local
pair_agreement 0.4731—0 ms1—
se_random_n100_pick_v1measuredsummeval_100—
random · local
top1 0.0701—0 ms1—
se_random_n100_v1measuredsummeval_100—
random · local
tau 0.0031—0 ms1—
se_rouge1_n100_v1measuredsummeval_100—
rouge · local
tau 0.1491—196 ms1—
se_rouge2_n100_v1measuredsummeval_100—
rouge · local
tau 0.1581—181 ms1—
se_rougeL_n100_v1measuredsummeval_100—
rouge · local
tau 0.1431—274 ms1—
se_unieval_n100_v1measuredsummeval_100MingZhong/unieval-sum
unieval · local
tau 0.3481—41676 ms1—
se_unievalpub_n100_v1measuredsummeval_100MingZhong/unieval-sum
unieval · local
tau 0.3721—44694 ms1—

02configs

One row per config, across every dataset it ran on: where it ranks on each goal is its best arm among the arms the goal's decision weighs.

267 shown
configmodel · providerdatasetsarmsspendrank on each goalruns in periodspend in period
ag_anthropic_l_n200_v1eval-opus-5
litellm
ag_news_2002$0.1750News topic classification (AG News) #5 of 301—
ag_anthropic_m_n200_v1eval-claude-sonnet
litellm
ag_news_2002$0.0703News topic classification (AG News) #4 of 301—
ag_anthropic_s_n200_v1eval-claude-haiku
litellm
ag_news_2002$0.0254News topic classification (AG News) #24 of 301—
ag_bart_mnli_n200_v1facebook/bart-large-mnli
hf_zeroshot · local
ag_news_2002—News topic classification (AG News) #28 of 301—
ag_bert_base_fy_n200_v1fabriceyhc/bert-base-uncased-ag_news
hf_local · local
ag_news_2003—News topic classification (AG News) #2 of 302—
ag_bert_base_ta_n200_v1textattack/bert-base-uncased-ag-news
hf_local · local
ag_news_2003—News topic classification (AG News) #1 of 302—
ag_bert_mini_n200_v1mrm8488/bert-mini-finetuned-age_news-classification
hf_local · local
ag_news_2003—News topic classification (AG News) #3 of 302—
ag_constant_n200_v1—
constant · local
ag_news_2002—News topic classification (AG News) #30 of 301—
ag_deepseek_l_n200_v1eval-deepseek-pro
litellm
ag_news_2002$0.0114News topic classification (AG News) #13 of 301—
ag_deepseek_m_n200_v1eval-deepseek-41flash
litellm
ag_news_2002$0.0043News topic classification (AG News) #23 of 301—
ag_deepseek_s_n200_v1eval-deepseek-flash
litellm
ag_news_2002$0.0014News topic classification (AG News) #25 of 301—
ag_gemma_l_n200_v1eval-gemma-31b
litellm
ag_news_2002$0.0028News topic classification (AG News) #14 of 301—
ag_gemma_m_n200_v1eval-gemma-26b
litellm
ag_news_2002$0.0016News topic classification (AG News) #21 of 301—
ag_gemma_s_n200_v1eval-gemma-3-27b
litellm
ag_news_2002$0.0021News topic classification (AG News) #10 of 301—
ag_glm_l_n200_v1eval-glm-5
litellm
ag_news_2002$0.0186News topic classification (AG News) #16 of 301—
ag_glm_m_n200_v1eval-glm-46
litellm
ag_news_2002$0.0107News topic classification (AG News) #17 of 301—
ag_glm_s_n200_v1eval-glm-45-air
litellm
ag_news_2002$0.0027News topic classification (AG News) #22 of 301—
ag_keyword_n200_v1—
keyword · local
ag_news_2002—News topic classification (AG News) #29 of 301—
ag_llama_l_n200_v1eval-llama-maverick
litellm
ag_news_2002$0.0043News topic classification (AG News) #26 of 301—
ag_llama_m_n200_v1eval-llama-70b
litellm
ag_news_2002$0.0040News topic classification (AG News) #9 of 301—
ag_llama_s_n200_v1eval-llama-scout
litellm
ag_news_2002$0.0025News topic classification (AG News) #27 of 301—
ag_mistral_l_n200_v1eval-mistral-large
litellm
ag_news_2002$0.0103News topic classification (AG News) #11 of 301—
ag_mistral_m_n200_v1eval-mistral-medium
litellm
ag_news_2002$0.0096News topic classification (AG News) #18 of 301—
ag_mistral_s_n200_v1eval-mistral-small
litellm
ag_news_2002$0.0019News topic classification (AG News) #19 of 301—
ag_openai_l_n200_v1eval-gpt-6sol
litellm
ag_news_2002$0.0487News topic classification (AG News) #7 of 301—
ag_openai_m_n200_v1eval-gpt-55
litellm
ag_news_2002$0.1272News topic classification (AG News) #6 of 301—
ag_openai_s_n200_v1eval-gpt-mini
litellm
ag_news_2002$0.0190News topic classification (AG News) #20 of 301—
ag_qwen_l_n200_v1eval-qwen-3max
litellm
ag_news_2002$0.0164News topic classification (AG News) #15 of 301—
ag_qwen_m_n200_v1eval-qwen-37plus
litellm
ag_news_2002$0.0070News topic classification (AG News) #8 of 301—
ag_qwen_s_n200_v1eval-qwen-flash
litellm
ag_news_2002$0.0036News topic classification (AG News) #12 of 301—
cnn_anthropic_l_n200_v1eval-opus-5
litellm
cnn_dailymail_2005$1.97Short news summaries (CNN) #18 of 291—
cnn_anthropic_m_n200_v1eval-claude-sonnet
litellm
cnn_dailymail_2005$0.7478Short news summaries (CNN) #15 of 291—
cnn_anthropic_s_n200_v1eval-claude-haiku
litellm
cnn_dailymail_2005$0.2611Short news summaries (CNN) #14 of 291—
cnn_bart_l_dm200_v1facebook/bart-large-cnn
hf_local · local
cnn_dailymail_dm2001——1—
cnn_bart_l_n200_v1facebook/bart-large-cnn
hf_local · local
cnn_dailymail_2007—Short news summaries (CNN) #1 of 292—
cnn_bart_l_v1facebook/bart-large-cnn
hf_local · local
cnn_dailymail_201——1—
cnn_bart_l_xsum_dm200_v1facebook/bart-large-xsum
hf_local · local
cnn_dailymail_dm2001——1—
cnn_bart_l_xsum_n200_v1facebook/bart-large-xsum
hf_local · local
cnn_dailymail_2006—Short news summaries (CNN) #29 of 292—
cnn_bart_m_dm200_v1sshleifer/distilbart-cnn-12-6
hf_local · local
cnn_dailymail_dm2001——1—
cnn_bart_m_n200_v1sshleifer/distilbart-cnn-12-6
hf_local · local
cnn_dailymail_2006—Short news summaries (CNN) #5 of 292—
cnn_bart_s_dm200_v1sshleifer/distilbart-cnn-6-6
hf_local · local
cnn_dailymail_dm2001——1—
cnn_bart_s_n200_v1sshleifer/distilbart-cnn-6-6
hf_local · local
cnn_dailymail_2007—Short news summaries (CNN) #10 of 293—
cnn_deepseek_l_n200_v1eval-deepseek-pro
litellm
cnn_dailymail_2005$0.1389Short news summaries (CNN) #7 of 291—
cnn_deepseek_m_n200_v1eval-deepseek-41flash
litellm
cnn_dailymail_2006$0.0381Short news summaries (CNN) #2 of 291—
cnn_deepseek_s_n200_v1eval-deepseek-flash
litellm
cnn_dailymail_2005$0.0137Short news summaries (CNN) #9 of 291—
cnn_gemma_l_n200_v1eval-gemma-31b
litellm
cnn_dailymail_2005$0.0237Short news summaries (CNN) #16 of 291—
cnn_gemma_m_n200_v1eval-gemma-26b
litellm
cnn_dailymail_2005$0.0154Short news summaries (CNN) #20 of 291—
cnn_gemma_s_n200_v1eval-gemma-3-27b
litellm
cnn_dailymail_2005$0.0200Short news summaries (CNN) #22 of 291—
cnn_glm_l_n200_v1eval-glm-5
litellm
cnn_dailymail_2005$0.1532Short news summaries (CNN) #21 of 291—
cnn_glm_m_n200_v1eval-glm-46
litellm
cnn_dailymail_2005$0.0995Short news summaries (CNN) #8 of 291—
cnn_glm_s_n200_v1eval-glm-45-air
litellm
cnn_dailymail_2005$0.0304Short news summaries (CNN) #19 of 291—
cnn_lead3_dm200_v1—
lead_k · local
cnn_dailymail_dm2001——1—
cnn_lead3_n200_v1—
lead_k · local
cnn_dailymail_2006—Short news summaries (CNN) #28 of 292—
cnn_lead3_v1—
lead_k · local
cnn_dailymail_201——1—
cnn_llama_l_n200_v1eval-llama-maverick
litellm
cnn_dailymail_2005$0.0389Short news summaries (CNN) #3 of 291—
cnn_llama_m_n200_v1eval-llama-70b
litellm
cnn_dailymail_2005$0.0330Short news summaries (CNN) #4 of 291—
cnn_llama_s_n200_v1eval-llama-scout
litellm
cnn_dailymail_2005$0.0222Short news summaries (CNN) #6 of 291—
cnn_mistral_l_n200_v1eval-mistral-large
litellm
cnn_dailymail_2005$0.0973Short news summaries (CNN) #24 of 291—
cnn_mistral_m_n200_v1eval-mistral-medium
litellm
cnn_dailymail_2005$0.0871Short news summaries (CNN) #26 of 291—
cnn_mistral_s_n200_v1eval-mistral-small
litellm
cnn_dailymail_2005$0.0179Short news summaries (CNN) #13 of 291—
cnn_openai_l_n200_v1eval-gpt-6sol
litellm
cnn_dailymail_2005$0.4362Short news summaries (CNN) #12 of 291—
cnn_openai_m_n200_v1eval-gpt-55
litellm
cnn_dailymail_2005$1.07Short news summaries (CNN) #11 of 291—
cnn_openai_s_n200_v1eval-gpt-mini
litellm
cnn_dailymail_2005$0.1672Short news summaries (CNN) #17 of 291—
cnn_qwen_l_n200_v1eval-qwen-3max
litellm
cnn_dailymail_2005$0.1689Short news summaries (CNN) #23 of 291—
cnn_qwen_m_n200_v1eval-qwen-37plus
litellm
cnn_dailymail_2005$0.0661Short news summaries (CNN) #27 of 291—
cnn_qwen_s_n200_v1eval-qwen-flash
litellm
cnn_dailymail_2005$0.0270Short news summaries (CNN) #25 of 291—
db_anthropic_l_n200_v1eval-opus-5
litellm
dbpedia_2801$0.3784Encyclopedia topic classification (DBpedia) #2 of 281—
db_anthropic_m_n200_v1eval-claude-sonnet
litellm
dbpedia_2801$0.1514Encyclopedia topic classification (DBpedia) #8 of 281—
db_anthropic_s_n200_v1eval-claude-haiku
litellm
dbpedia_2801$0.0536Encyclopedia topic classification (DBpedia) #10 of 281—
db_bart_mnli_n200_v1facebook/bart-large-mnli
hf_zeroshot · local
dbpedia_2801—Encyclopedia topic classification (DBpedia) #27 of 281—
db_bert_base_fy_n200_v1fabriceyhc/bert-base-uncased-dbpedia_14
hf_local · local
dbpedia_2802—Encyclopedia topic classification (DBpedia) #3 of 282—
db_constant_n200_v1—
constant · local
dbpedia_2801—Encyclopedia topic classification (DBpedia) #28 of 281—
db_deepseek_l_n200_v1eval-deepseek-pro
litellm
dbpedia_2801$0.0209Encyclopedia topic classification (DBpedia) #4 of 281—
db_deepseek_m_n200_v1eval-deepseek-41flash
litellm
dbpedia_2801$0.0079Encyclopedia topic classification (DBpedia) #12 of 281—
db_deepseek_s_n200_v1eval-deepseek-flash
litellm
dbpedia_2801$0.0028Encyclopedia topic classification (DBpedia) #24 of 281—
db_gemma_l_n200_v1eval-gemma-31b
litellm
dbpedia_2801$0.0046Encyclopedia topic classification (DBpedia) #13 of 281—
db_gemma_m_n200_v1eval-gemma-26b
litellm
dbpedia_2801$0.0033Encyclopedia topic classification (DBpedia) #5 of 281—
db_gemma_s_n200_v1eval-gemma-3-27b
litellm
dbpedia_2801$0.0035Encyclopedia topic classification (DBpedia) #14 of 281—
db_glm_l_n200_v1eval-glm-5
litellm
dbpedia_2801$0.0359Encyclopedia topic classification (DBpedia) #18 of 281—
db_glm_m_n200_v1eval-glm-46
litellm
dbpedia_2801$0.0159Encyclopedia topic classification (DBpedia) #6 of 281—
db_glm_s_n200_v1eval-glm-45-air
litellm
dbpedia_2801$0.0052Encyclopedia topic classification (DBpedia) #25 of 281—
db_keyword_n200_v1—
keyword · local
dbpedia_2801—Encyclopedia topic classification (DBpedia) #26 of 281—
db_llama_l_n200_v1eval-llama-maverick
litellm
dbpedia_2801$0.0089Encyclopedia topic classification (DBpedia) #15 of 281—
db_llama_m_n200_v1eval-llama-70b
litellm
dbpedia_2801$0.0084Encyclopedia topic classification (DBpedia) #11 of 281—
db_llama_s_n200_v1eval-llama-scout
litellm
dbpedia_2801$0.0076Encyclopedia topic classification (DBpedia) #21 of 281—
db_mistral_l_n200_v1eval-mistral-large
litellm
dbpedia_2801$0.0210Encyclopedia topic classification (DBpedia) #20 of 281—
db_mistral_m_n200_v1eval-mistral-medium
litellm
dbpedia_2801$0.0188Encyclopedia topic classification (DBpedia) #22 of 281—
db_mistral_s_n200_v1eval-mistral-small
litellm
dbpedia_2801$0.0039Encyclopedia topic classification (DBpedia) #23 of 281—
db_openai_l_n200_v1eval-gpt-6sol
litellm
dbpedia_2801$0.0921Encyclopedia topic classification (DBpedia) #16 of 281—
db_openai_m_n200_v1eval-gpt-55
litellm
dbpedia_2801$0.2380Encyclopedia topic classification (DBpedia) #9 of 281—
db_openai_s_n200_v1eval-gpt-mini
litellm
dbpedia_2801$0.0360Encyclopedia topic classification (DBpedia) #17 of 281—
db_qwen_l_n200_v1eval-qwen-3max
litellm
dbpedia_2801$0.0342Encyclopedia topic classification (DBpedia) #19 of 281—
db_qwen_m_n200_v1eval-qwen-37plus
litellm
dbpedia_2801$0.0143Encyclopedia topic classification (DBpedia) #1 of 281—
db_qwen_s_n200_v1eval-qwen-flash
litellm
dbpedia_2801$0.0072Encyclopedia topic classification (DBpedia) #7 of 281—
dedup_guard_v1—
— · local
fleet_dedup_v11——1—
dedup_pinned_f07efa9_v1—
— · local
fleet_dedup_v11——1—
demo_v1—
echo · local
smoke_v11——1—
fn_anthropic_l_n200_v1eval-opus-5
litellm
few_nerd_2801$0.6897Named-entity extraction (Few-NERD) #3 of 281—
fn_anthropic_m_n200_v1eval-claude-sonnet
litellm
few_nerd_2801$0.2846Named-entity extraction (Few-NERD) #6 of 281—
fn_anthropic_s_n200_v1eval-claude-haiku
litellm
few_nerd_2801$0.1552Named-entity extraction (Few-NERD) #18 of 281—
fn_capitalized_n200_v1—
capitalized · local
few_nerd_2801—Named-entity extraction (Few-NERD) #27 of 281—
fn_deepseek_l_n200_v1eval-deepseek-pro
litellm
few_nerd_2801$0.0853Named-entity extraction (Few-NERD) #17 of 281—
fn_deepseek_m_n200_v1eval-deepseek-41flash
litellm
few_nerd_2801$0.0200Named-entity extraction (Few-NERD) #11 of 281—
fn_deepseek_s_n200_v1eval-deepseek-flash
litellm
few_nerd_2801$0.0091Named-entity extraction (Few-NERD) #19 of 281—
fn_gemma_l_n200_v1eval-gemma-31b
litellm
few_nerd_2801$0.0134Named-entity extraction (Few-NERD) #9 of 281—
fn_gemma_m_n200_v1eval-gemma-26b
litellm
few_nerd_2801$0.0088Named-entity extraction (Few-NERD) #4 of 281—
fn_gemma_s_n200_v1eval-gemma-3-27b
litellm
few_nerd_2801$0.0093Named-entity extraction (Few-NERD) #24 of 281—
fn_gliner_n200_v1urchade/gliner_medium-v2.1
gliner · local
few_nerd_2801—Named-entity extraction (Few-NERD) #26 of 281—
fn_glm_l_n200_v1eval-glm-5
litellm
few_nerd_2801$0.1023Named-entity extraction (Few-NERD) #22 of 281—
fn_glm_m_n200_v1eval-glm-46
litellm
few_nerd_2801$0.0445Named-entity extraction (Few-NERD) #15 of 281—
fn_glm_s_n200_v1eval-glm-45-air
litellm
few_nerd_2801$0.0194Named-entity extraction (Few-NERD) #23 of 281—
fn_llama_l_n200_v1eval-llama-maverick
litellm
few_nerd_2801$0.0176Named-entity extraction (Few-NERD) #12 of 281—
fn_llama_m_n200_v1eval-llama-70b
litellm
few_nerd_2801$0.0194Named-entity extraction (Few-NERD) #20 of 281—
fn_llama_s_n200_v1eval-llama-scout
litellm
few_nerd_2801$0.0103Named-entity extraction (Few-NERD) #21 of 281—
fn_mistral_l_n200_v1eval-mistral-large
litellm
few_nerd_2801$0.0419Named-entity extraction (Few-NERD) #13 of 281—
fn_mistral_m_n200_v1eval-mistral-medium
litellm
few_nerd_2801$0.0462Named-entity extraction (Few-NERD) #16 of 281—
fn_mistral_s_n200_v1eval-mistral-small
litellm
few_nerd_2801$0.0076Named-entity extraction (Few-NERD) #25 of 281—
fn_nothing_n200_v1—
nothing · local
few_nerd_2801—Named-entity extraction (Few-NERD) #28 of 281—
fn_openai_l_n200_v1eval-gpt-6sol
litellm
few_nerd_2801$0.1715Named-entity extraction (Few-NERD) #5 of 281—
fn_openai_m_n200_v1eval-gpt-55
litellm
few_nerd_2801$0.6507Named-entity extraction (Few-NERD) #2 of 281—
fn_openai_s_n200_v1eval-gpt-mini
litellm
few_nerd_2801$0.0784Named-entity extraction (Few-NERD) #14 of 281—
fn_qwen_l_n200_v1eval-qwen-3max
litellm
few_nerd_2801$0.0722Named-entity extraction (Few-NERD) #10 of 281—
fn_qwen_m_n200_v1eval-qwen-37plus
litellm
few_nerd_2801$0.0396Named-entity extraction (Few-NERD) #8 of 281—
fn_qwen_s_n200_v1eval-qwen-flash
litellm
few_nerd_2801$0.0157Named-entity extraction (Few-NERD) #7 of 281—
fn_span_marker_n200_v1guishe/span-marker-generic-ner-v1-fewnerd-fine-super
span_marker · local
few_nerd_2801—Named-entity extraction (Few-NERD) #1 of 281—
se_anthropic_l_n100_v1eval-opus-5
litellm
summeval_1001$2.01A scorer that agrees with experts on summaries (SummEval) #3 of 331—
se_anthropic_m_n100_v1eval-claude-sonnet
litellm
summeval_1001$0.8048A scorer that agrees with experts on summaries (SummEval) #12 of 331—
se_anthropic_s_n100_v1eval-claude-haiku
litellm
summeval_1001$0.3237A scorer that agrees with experts on summaries (SummEval) #11 of 331—
se_bartscore_n100_v1facebook/bart-large-cnn
bartscore · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #27 of 331—
se_coverage_n100_v1—
coverage · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #31 of 331—
se_deepseek_l_n100_v1eval-deepseek-pro
litellm
summeval_1001$0.0258A scorer that agrees with experts on summaries (SummEval) #14 of 331—
se_deepseek_m_n100_v1eval-deepseek-41flash
litellm
summeval_1001$0.0771A scorer that agrees with experts on summaries (SummEval) #10 of 331—
se_deepseek_s_n100_v1eval-deepseek-flash
litellm
summeval_1001$0.0307A scorer that agrees with experts on summaries (SummEval) #16 of 331—
se_gemma_l_n100_pairwise_v1eval-gemma-31b
litellm
summeval_1001$0.0764—1—
se_gemma_l_n100_pick_v1eval-gemma-31b
litellm
summeval_1001$0.0209—1—
se_gemma_l_n100_v1eval-gemma-31b
litellm
summeval_1001$0.0278A scorer that agrees with experts on summaries (SummEval) #2 of 331—
se_gemma_l_pubtext_n100_v1eval-gemma-31b
litellm
summeval_1001$0.1013—1—
se_gemma_l_pubtextcw_n100_v1eval-gemma-31b
litellm
summeval_1001$0.0276—1—
se_gemma_m_n100_v1eval-gemma-26b
litellm
summeval_1001$0.0252A scorer that agrees with experts on summaries (SummEval) #13 of 331—
se_gemma_s_n100_v1eval-gemma-3-27b
litellm
summeval_1001$0.0279A scorer that agrees with experts on summaries (SummEval) #21 of 331—
se_glm_l_n100_v1eval-glm-5
litellm
summeval_1001$0.2451A scorer that agrees with experts on summaries (SummEval) #9 of 331—
se_glm_m_n100_v1eval-glm-46
litellm
summeval_1001$0.1144A scorer that agrees with experts on summaries (SummEval) #7 of 331—
se_glm_s_n100_v1eval-glm-45-air
litellm
summeval_1001$0.0481A scorer that agrees with experts on summaries (SummEval) #19 of 331—
se_length_n100_v1—
length · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #32 of 331—
se_llama_l_n100_v1eval-llama-maverick
litellm
summeval_1001$0.0594A scorer that agrees with experts on summaries (SummEval) #18 of 331—
se_llama_m_n100_v1eval-llama-70b
litellm
summeval_1001$0.0517A scorer that agrees with experts on summaries (SummEval) #23 of 331—
se_llama_s_n100_v1eval-llama-scout
litellm
summeval_1001$0.0297A scorer that agrees with experts on summaries (SummEval) #25 of 331—
se_mistral_l_n100_v1eval-mistral-large
litellm
summeval_1001$0.1257A scorer that agrees with experts on summaries (SummEval) #17 of 331—
se_mistral_m_n100_v1eval-mistral-medium
litellm
summeval_1001$0.1191A scorer that agrees with experts on summaries (SummEval) #15 of 331—
se_mistral_s_n100_v1eval-mistral-small
litellm
summeval_1001$0.0283A scorer that agrees with experts on summaries (SummEval) #20 of 331—
se_openai_l_n100_v1eval-gpt-6sol
litellm
summeval_1001$0.6350A scorer that agrees with experts on summaries (SummEval) #1 of 331—
se_openai_m_n100_v1eval-gpt-55
litellm
summeval_1001$1.42A scorer that agrees with experts on summaries (SummEval) #4 of 331—
se_openai_s_n100_v1eval-gpt-mini
litellm
summeval_1001$0.2124A scorer that agrees with experts on summaries (SummEval) #22 of 331—
se_qwen37max_n100_v1eval-qwen-37max
litellm
summeval_1001$0.4436A scorer that agrees with experts on summaries (SummEval) #5 of 331—
se_qwen_m_n100_v1eval-qwen-37plus
litellm
summeval_1001$0.1084A scorer that agrees with experts on summaries (SummEval) #8 of 331—
se_qwen_s_n100_pairwise_v1eval-qwen-flash
litellm
summeval_1001$0.0531—1—
se_qwen_s_n100_pick_v1eval-qwen-flash
litellm
summeval_1001$0.0282—1—
se_qwen_s_n100_v1eval-qwen-flash
litellm
summeval_1001$0.0385A scorer that agrees with experts on summaries (SummEval) #6 of 331—
se_random_n100_pairwise_v1—
random · local
summeval_1001——1—
se_random_n100_pick_v1—
random · local
summeval_1001——1—
se_random_n100_v1—
random · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #33 of 331—
se_rouge1_n100_v1—
rouge · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #29 of 331—
se_rouge2_n100_v1—
rouge · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #28 of 331—
se_rougeL_n100_v1—
rouge · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #30 of 331—
se_unieval_n100_v1MingZhong/unieval-sum
unieval · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #26 of 331—
se_unievalpub_n100_v1MingZhong/unieval-sum
unieval · local
summeval_1001—A scorer that agrees with experts on summaries (SummEval) #24 of 331—
sf_bge_small_n200_v1BAAI/bge-small-en-v1.5
dense · local
scifact_2001—Evidence retrieval for scientific claims (SciFact) #11 of 191—
sf_bm25_n200_v1—
bm25 · local
scifact_2001—Evidence retrieval for scientific claims (SciFact) #17 of 191—
sf_deepseek_m_n200_v1eval-deepseek-41flash
rerank
scifact_2001$0.2381Evidence retrieval for scientific claims (SciFact) #3 of 191—
sf_deepseek_s_n200_v1eval-deepseek-flash
rerank
scifact_2001$0.0600Evidence retrieval for scientific claims (SciFact) #4 of 191—
sf_e5_base_n200_v1intfloat/e5-base-v2
dense · local
scifact_2001—Evidence retrieval for scientific claims (SciFact) #8 of 191—
sf_first_k_n200_v1—
first_k · local
scifact_2001—Evidence retrieval for scientific claims (SciFact) #19 of 191—
sf_gemma_l_n200_v1eval-gemma-31b
rerank
scifact_2001$0.1701Evidence retrieval for scientific claims (SciFact) #6 of 191—
sf_gemma_m_n200_v1eval-gemma-26b
rerank
scifact_2001$0.0721Evidence retrieval for scientific claims (SciFact) #5 of 191—
sf_gemma_s_n200_v1eval-gemma-3-27b
rerank
scifact_2001$0.0962Evidence retrieval for scientific claims (SciFact) #14 of 191—
sf_glm_m_n200_v1eval-glm-46
rerank
scifact_2001$0.4576Evidence retrieval for scientific claims (SciFact) #2 of 191—
sf_glm_s_bm25first_n200_v1eval-glm-45-air
rerank
scifact_2001$0.1481—1—
sf_glm_s_e5first_n200_v1eval-glm-45-air
rerank
scifact_2001$0.1464—1—
sf_glm_s_n200_v1eval-glm-45-air
rerank
scifact_2001$0.1421Evidence retrieval for scientific claims (SciFact) #1 of 191—
sf_llama_l_n200_v1eval-llama-maverick
rerank
scifact_2002$0.3422Evidence retrieval for scientific claims (SciFact) #7 of 192—
sf_llama_m_n200_v1eval-llama-70b
rerank
scifact_2001$0.1434Evidence retrieval for scientific claims (SciFact) #12 of 191—
sf_llama_s_n200_v1eval-llama-scout
rerank
scifact_2002$0.2506Evidence retrieval for scientific claims (SciFact) #10 of 192—
sf_minilm_n200_v1sentence-transformers/all-MiniLM-L6-v2
dense · local
scifact_2001—Evidence retrieval for scientific claims (SciFact) #16 of 191—
sf_mistral_s_n200_v1eval-mistral-small
rerank
scifact_2001$0.0852Evidence retrieval for scientific claims (SciFact) #13 of 191—
sf_mpnet_n200_v1sentence-transformers/all-mpnet-base-v2
dense · local
scifact_2001—Evidence retrieval for scientific claims (SciFact) #15 of 191—
sf_qwen_s_bm25first_n200_v1eval-qwen-flash
rerank
scifact_2001$0.1551—1—
sf_qwen_s_e5first_n200_v1eval-qwen-flash
rerank
scifact_2001$0.1508—1—
sf_qwen_s_n200_v1eval-qwen-flash
rerank
scifact_2002$0.3117Evidence retrieval for scientific claims (SciFact) #9 of 192—
sf_random_n200_v1—
random · local
scifact_2001—Evidence retrieval for scientific claims (SciFact) #18 of 191—
triage_baseline_pro_noreason_v1eval-fleet-triage-pro
gateway
orrery_triage_v1_labelled3$0.3160Triage decisions on Orrery's error signals #52 of 551—
triage_baseline_pro_v1eval-fleet-triage-pro
gateway
orrery_triage_v1_labelled10$1.32Triage decisions on Orrery's error signals #35 of 554—
triage_deepseek_41flash_v1eval-deepseek-41flash
gateway
orrery_triage_v1_labelled4$0.0560—1—
triage_deepseek_41flash_v2eval-deepseek-41flash
gateway
orrery_triage_v1_labelled3$0.0231Triage decisions on Orrery's error signals #47 of 551—
triage_gemini_35_flash_lite_v1eval-gemini-35-flash-lite
gateway
orrery_triage_v1_labelled5$0.0889Triage decisions on Orrery's error signals #6 of 551—
triage_gemini_4_argon_v1eval-gemini-4-argon
gateway
orrery_triage_v1_labelled5$0.8851Triage decisions on Orrery's error signals #17 of 551—
triage_gpt_61_sol_v1eval-gpt-61-sol
gateway
orrery_triage_v1_labelled10$2.83Triage decisions on Orrery's error signals #9 of 554—
triage_gpt_6_luna_v1eval-gpt-6-luna
gateway
orrery_triage_v1_labelled8$0.1798Triage decisions on Orrery's error signals #15 of 554—
triage_gpt_6_luna_v2eval-gpt-6-luna
gateway
orrery_triage_v1_labelled3$0.0434Triage decisions on Orrery's error signals #7 of 551—
triage_haiku_55_v1eval-claude-haiku-55
gateway
orrery_triage_v1_labelled10$0.2646Triage decisions on Orrery's error signals #10 of 554—
triage_hosted_gemma4_31b_v1eval-gemma-31b
gateway
orrery_triage_v1_labelled4$0.0587—1—
triage_hosted_gemma4_31b_v2eval-gemma-31b
gateway
orrery_triage_v1_labelled3$0.0583Triage decisions on Orrery's error signals #29 of 551—
triage_hosted_gpt_oss_120b_v1eval-gpt-oss-120b
gateway
orrery_triage_v1_labelled4$0.0552—1—
triage_hosted_gpt_oss_120b_v2eval-gpt-oss-120b
gateway
orrery_triage_v1_labelled3$0.0407Triage decisions on Orrery's error signals #16 of 551—
triage_hosted_mistral_medium_35_v1eval-mistral-medium-35
gateway
orrery_triage_v1_labelled5$0.7500Triage decisions on Orrery's error signals #44 of 551—
triage_hosted_qwen35_122b_v1eval-qwen35-122b
gateway
orrery_triage_v1_labelled4$0.2177—1—
triage_hosted_qwen35_122b_v2eval-qwen35-122b
gateway
orrery_triage_v1_labelled3$0.1075Triage decisions on Orrery's error signals #45 of 551—
triage_hosted_qwen3_30b_a3b_v1eval-qwen3-30b-a3b
gateway
orrery_triage_v1_labelled5$0.0455Triage decisions on Orrery's error signals #25 of 551—
triage_kimi_k31_pro_v1eval-kimi-k31-pro
gateway
orrery_triage_v1_labelled4$1.24—1—
triage_kimi_k31_pro_v2eval-kimi-k31-pro
gateway
orrery_triage_v1_labelled3$0.7826Triage decisions on Orrery's error signals #30 of 551—
triage_local_dgx_gemma4_12b_noreason_v1gemma4:12b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #11 of 551—
triage_local_dgx_gemma4_31b_v1gemma4:31b
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_gemma4_31b_v2gemma4:31b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #31 of 551—
triage_local_dgx_gpt_oss_120b_v1gpt-oss:120b
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_gpt_oss_120b_v2gpt-oss:120b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #19 of 551—
triage_local_dgx_gpt_oss_20b_v1gpt-oss:20b
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_gpt_oss_20b_v2gpt-oss:20b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #32 of 551—
triage_local_dgx_llama31_8b_v1llama3.1:8b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #2 of 551—
triage_local_dgx_llama4_scout_v1llama4:scout
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #53 of 551—
triage_local_dgx_ministral_3_14b_v1ministral-3:14b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #41 of 551—
triage_local_dgx_mistral_medium_35_128b_v1mistral-medium-3.5:128b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #49 of 551—
triage_local_dgx_nemotron_35_bolt_e4b_v1nemotron-3.5-bolt:e4b-q4_K_M
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_nemotron_35_bolt_e4b_v2nemotron-3.5-bolt:e4b-q4_K_M
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #18 of 551—
triage_local_dgx_nemotron_35_lightning_30b_v1nemotron-3.5-lightning:30b
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_nemotron_35_lightning_30b_v2nemotron-3.5-lightning:30b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #21 of 551—
triage_local_dgx_nemotron_35_super_120b_v1nemotron-3.5-super:120b
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_nemotron_35_super_120b_v2nemotron-3.5-super:120b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #22 of 551—
triage_local_dgx_phi4_14b_v1phi4:14b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #20 of 551—
triage_local_dgx_qwen35_122b_v1qwen3.5:122b
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_qwen35_122b_v2qwen3.5:122b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #55 of 551—
triage_local_dgx_qwen38_27b_v1qwen3.8:27b
gateway
orrery_triage_v1_labelled4——1—
triage_local_dgx_qwen38_27b_v2qwen3.8:27b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #50 of 551—
triage_local_dgx_qwen3_30b_a3b_v1qwen3:30b-a3b-instruct-2507-q4_K_M
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #46 of 551—
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1NVFP4/Qwen3-30B-A3B-Instruct-2507-FP4
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #1 of 551—
triage_local_mac_gemma4_12b_noreason_v1gemma4:12b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #12 of 551—
triage_local_mac_gemma4_31b_v1gemma4:31b-it-q4_K_M
gateway
orrery_triage_v1_labelled4——1—
triage_local_mac_gemma4_31b_v2gemma4:31b-it-q4_K_M
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #37 of 551—
triage_local_mac_gpt_oss_20b_v1gpt-oss:20b
gateway
orrery_triage_v1_labelled4——1—
triage_local_mac_gpt_oss_20b_v2gpt-oss:20b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #26 of 551—
triage_local_mac_llama31_8b_v1llama3.1:8b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #3 of 551—
triage_local_mac_ministral_3_14b_v1ministral-3:14b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #48 of 551—
triage_local_mac_nemotron_35_bolt_e4b_v1nemotron-3.5-bolt:e4b-q4_K_M
gateway
orrery_triage_v1_labelled4——1—
triage_local_mac_nemotron_35_bolt_e4b_v2nemotron-3.5-bolt:e4b-q4_K_M
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #13 of 551—
triage_local_mac_nemotron_35_lightning_30b_v1nemotron-3.5-lightning:30b-a3b-q4_K_M
gateway
orrery_triage_v1_labelled4——1—
triage_local_mac_nemotron_35_lightning_30b_v2nemotron-3.5-lightning:30b-a3b-q4_K_M
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #24 of 551—
triage_local_mac_phi4_14b_v1phi4:14b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #34 of 551—
triage_local_mac_qwen38_27b_v1qwen3.8:27b
gateway
orrery_triage_v1_labelled4——1—
triage_local_mac_qwen38_27b_v2qwen3.8:27b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #51 of 551—
triage_local_mac_qwen3_30b_a3b_v1qwen3:30b-a3b-instruct-2507-q4_K_M
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #42 of 551—
triage_local_x3_gemma4_31b_v1gemma4:31b
gateway
orrery_triage_v1_labelled4——1—
triage_local_x3_gemma4_31b_v2gemma4:31b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #27 of 551—
triage_local_x3_gpt_oss_20b_v1gpt-oss:20b
gateway
orrery_triage_v1_labelled4——1—
triage_local_x3_gpt_oss_20b_v2gpt-oss:20b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #39 of 551—
triage_local_x3_llama31_8b_v1llama3.1:8b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #4 of 551—
triage_local_x3_ministral_3_14b_v1ministral-3:14b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #40 of 551—
triage_local_x3_nemotron_35_bolt_e4b_v1nemotron-3.5-bolt:e4b-q4_K_M
gateway
orrery_triage_v1_labelled8—Triage decisions on Orrery's error signals #14 of 552—
triage_local_x3_nemotron_35_bolt_e4b_v2nemotron-3.5-bolt:e4b-q4_K_M
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #8 of 551—
triage_local_x3_nemotron_35_lightning_30b_v1nemotron-3.5-lightning:30b
gateway
orrery_triage_v1_labelled4——1—
triage_local_x3_nemotron_35_lightning_30b_v2nemotron-3.5-lightning:30b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #23 of 551—
triage_local_x3_phi4_14b_v1phi4:14b
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #36 of 551—
triage_local_x3_qwen38_27b_v1qwen3.8:27b
gateway
orrery_triage_v1_labelled4——1—
triage_local_x3_qwen38_27b_v2qwen3.8:27b
gateway
orrery_triage_v1_labelled3—Triage decisions on Orrery's error signals #54 of 551—
triage_local_x3_qwen3_30b_a3b_v1qwen3:30b-a3b-instruct-2507-q4_K_M
gateway
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #33 of 551—
triage_mistral_large_4_v1eval-mistral-large-4
gateway
orrery_triage_v1_labelled4$0.4464—1—
triage_mistral_large_4_v2eval-mistral-large-4
gateway
orrery_triage_v1_labelled3$0.4516Triage decisions on Orrery's error signals #38 of 551—
triage_sonnet_55_v1eval-claude-sonnet-55
gateway
orrery_triage_v1_labelled5$1.47Triage decisions on Orrery's error signals #28 of 551—
triage_stub_dismiss_v1—
stub · local
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #43 of 551—
triage_stub_escalate_v1—
stub · local
orrery_triage_v1_labelled5—Triage decisions on Orrery's error signals #5 of 551—

03what changed in the period

whenarmwhat
2026-11-0229 arms on cnn_dailymail_200judged by judge-summ-pick-v1
2026-11-0283 arms on orrery_triage_v1_labelledrescored against gold/orrery_triage_v1_labelled
2026-11-02se_gemma_l_n100_pairwise_v1measured
2026-11-02se_gemma_l_n100_pick_v1measured
2026-11-02se_qwen_s_n100_pairwise_v1measured
2026-11-02se_qwen_s_n100_pick_v1measured
2026-11-02se_random_n100_pairwise_v1measured
2026-11-02se_random_n100_pick_v1measured
2026-11-0156 arms on orrery_triage_v1_labelledrescored against gold/orrery_triage_v1_labelled
2026-11-01triage_local_mac_gemma4_31b_v2measured
2026-11-01triage_deepseek_41flash_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_gpt_6_luna_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_hosted_gemma4_31b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_hosted_gpt_oss_120b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_hosted_qwen35_122b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_kimi_k31_pro_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_gemma4_31b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_gpt_oss_120b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_gpt_oss_20b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_nemotron_35_bolt_e4b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_nemotron_35_lightning_30b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_nemotron_35_super_120b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_qwen35_122b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_dgx_qwen38_27b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_mac_gemma4_31b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_mac_gpt_oss_20b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_mac_nemotron_35_bolt_e4b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_mac_nemotron_35_lightning_30b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_mac_qwen38_27b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_x3_gemma4_31b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_x3_gpt_oss_20b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_x3_nemotron_35_bolt_e4b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_x3_nemotron_35_lightning_30b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_local_x3_qwen38_27b_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-11-01triage_mistral_large_4_v1retired in r3: run at the server's default reasoning for a model that reasons by default; re-run in this round as _v2 with the model's onboarded settings
2026-10-31triage_baseline_pro_noreason_v1measured
2026-10-31triage_deepseek_41flash_v2measured
2026-10-31triage_gpt_6_luna_v2measured
2026-10-31triage_hosted_gemma4_31b_v2measured
2026-10-31triage_hosted_gpt_oss_120b_v2measured
2026-10-31triage_hosted_qwen35_122b_v2measured
2026-10-31triage_kimi_k31_pro_v2measured
2026-10-31triage_local_dgx_gemma4_12b_noreason_v1measured
2026-10-31triage_local_dgx_gemma4_31b_v1measured
2026-10-31triage_local_dgx_gemma4_31b_v2measured
2026-10-31triage_local_dgx_gpt_oss_120b_v1measured
2026-10-31triage_local_dgx_gpt_oss_120b_v2measured
2026-10-31triage_local_dgx_gpt_oss_20b_v2measured
2026-10-31triage_local_dgx_llama31_8b_v1measured
2026-10-31triage_local_dgx_ministral_3_14b_v1measured
2026-10-31triage_local_dgx_nemotron_35_bolt_e4b_v2measured
2026-10-31triage_local_dgx_nemotron_35_lightning_30b_v2measured
2026-10-31triage_local_dgx_nemotron_35_super_120b_v2measured
2026-10-31triage_local_dgx_phi4_14b_v1measured
2026-10-31triage_local_dgx_qwen35_122b_v2measured
2026-10-31triage_local_dgx_qwen38_27b_v2measured
2026-10-31triage_local_dgx_qwen3_30b_a3b_v1measured
2026-10-31triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1measured
2026-10-31triage_local_mac_gemma4_12b_noreason_v1measured
2026-10-31triage_local_mac_gpt_oss_20b_v2measured
2026-10-31triage_local_mac_ministral_3_14b_v1measured
2026-10-31triage_local_mac_nemotron_35_bolt_e4b_v2measured
2026-10-31triage_local_mac_nemotron_35_lightning_30b_v2measured
2026-10-31triage_local_mac_qwen38_27b_v2measured
2026-10-31triage_local_mac_qwen3_30b_a3b_v1measured
2026-10-31triage_local_x3_gemma4_31b_v2measured
2026-10-31triage_local_x3_gpt_oss_20b_v2measured
2026-10-31triage_local_x3_nemotron_35_bolt_e4b_v2measured
2026-10-31triage_local_x3_nemotron_35_lightning_30b_v2measured
2026-10-31triage_local_x3_qwen38_27b_v2measured
2026-10-31triage_local_x3_qwen3_30b_a3b_v1measured
2026-10-31triage_mistral_large_4_v2measured
2026-10-31triage_stub_dismiss_v1measured
2026-10-31triage_stub_escalate_v1measured
2026-10-31114 arms on orrery_triage_v1_labelledrescored against gold/orrery_triage_v1_labelled
2026-10-31triage_baseline_pro_v1promoted as the baseline
2026-10-30dedup_guard_v1measured
2026-10-30dedup_pinned_f07efa9_v1measured
2026-10-30triage_baseline_pro_v1measured
2026-10-30triage_deepseek_41flash_v1measured
2026-10-30triage_gemini_35_flash_lite_v1measured
2026-10-30triage_gemini_4_argon_v1measured
2026-10-30triage_gpt_61_sol_v1measured
2026-10-30triage_gpt_6_luna_v1measured
2026-10-30triage_haiku_55_v1measured
2026-10-30triage_hosted_gemma4_31b_v1measured
2026-10-30triage_hosted_gpt_oss_120b_v1measured
2026-10-30triage_hosted_mistral_medium_35_v1measured
2026-10-30triage_hosted_qwen35_122b_v1measured
2026-10-30triage_hosted_qwen3_30b_a3b_v1measured
2026-10-30triage_kimi_k31_pro_v1measured
2026-10-30triage_local_dgx_gpt_oss_20b_v1measured
2026-10-30triage_local_dgx_llama4_scout_v1measured
2026-10-30triage_local_dgx_mistral_medium_35_128b_v1measured
2026-10-30triage_local_dgx_nemotron_35_bolt_e4b_v1measured
2026-10-30triage_local_dgx_nemotron_35_lightning_30b_v1measured
2026-10-30triage_local_dgx_nemotron_35_super_120b_v1measured
2026-10-30triage_local_dgx_qwen35_122b_v1measured
2026-10-30triage_local_dgx_qwen38_27b_v1measured
2026-10-30triage_local_mac_gemma4_31b_v1measured
2026-10-30triage_local_mac_gpt_oss_20b_v1measured
2026-10-30triage_local_mac_llama31_8b_v1measured
2026-10-30triage_local_mac_nemotron_35_bolt_e4b_v1measured
2026-10-30triage_local_mac_nemotron_35_lightning_30b_v1measured
2026-10-30triage_local_mac_phi4_14b_v1measured
2026-10-30triage_local_mac_qwen38_27b_v1measured
2026-10-30triage_local_x3_gemma4_31b_v1measured
2026-10-30triage_local_x3_gpt_oss_20b_v1measured
2026-10-30triage_local_x3_llama31_8b_v1measured
2026-10-30triage_local_x3_ministral_3_14b_v1measured
2026-10-30triage_local_x3_nemotron_35_bolt_e4b_v1measured
2026-10-30triage_local_x3_nemotron_35_lightning_30b_v1measured
2026-10-30triage_local_x3_phi4_14b_v1measured
2026-10-30triage_local_x3_qwen38_27b_v1measured
2026-10-30triage_mistral_large_4_v1measured
2026-10-30triage_sonnet_55_v1measured
2026-10-305 arms on orrery_triage_v1_labelledrun again (repeats)
2026-10-2530 arms on ag_news_200rescored against gold/ag_news_200_adj1
2026-10-2529 arms on cnn_dailymail_200judged by judge-summ-v1
2026-10-2529 arms on cnn_dailymail_200judged by judge-summ-v2
2026-10-2529 arms on cnn_dailymail_200rescored against silver/cnn_dailymail_200
2026-10-25cnn_bart_l_n200_v1 · judged/judge-summ-pointwise-v1judged by judge-summ-pointwise-v1
2026-10-25cnn_deepseek_m_n200_v1 · judged/judge-summ-pointwise-v1judged by judge-summ-pointwise-v1
2026-10-25ag_bert_mini_n200_v1promoted as the baseline
2026-10-25cnn_bart_l_n200_v1promoted as the baseline
2026-10-25db_bert_base_fy_n200_v1promoted as the baseline
2026-10-25fn_span_marker_n200_v1promoted as the baseline
2026-10-25se_gemma_l_n100_v1promoted as the baseline
2026-10-25sf_e5_base_n200_v1promoted as the baseline
2026-10-23demo_v1measured
2026-10-23demo_v1promoted as the baseline
2026-10-19se_gemma_l_pubtext_n100_v1 · pairrun again (pair)
2026-10-19se_gemma_l_pubtextcw_n100_v1 · pairrun again (pair)
2026-10-19se_gemma_l_pubtext_n100_v1 · pairretired in r2: a paired side run, not on the leaderboard; data/runs is the record
2026-10-19se_gemma_l_pubtextcw_n100_v1 · pairretired in r2: a paired side run, not on the leaderboard; data/runs is the record
2026-10-18se_anthropic_l_n100_v1measured
2026-10-18se_anthropic_m_n100_v1measured
2026-10-18se_anthropic_s_n100_v1measured
2026-10-18se_deepseek_l_n100_v1measured
2026-10-18se_deepseek_m_n100_v1measured
2026-10-18se_deepseek_s_n100_v1measured
2026-10-18se_gemma_l_n100_v1measured
2026-10-18se_gemma_m_n100_v1measured
2026-10-18se_gemma_s_n100_v1measured
2026-10-18se_glm_l_n100_v1measured
2026-10-18se_glm_m_n100_v1measured
2026-10-18se_glm_s_n100_v1measured
2026-10-18se_llama_l_n100_v1measured
2026-10-18se_llama_m_n100_v1measured
2026-10-18se_llama_s_n100_v1measured
2026-10-18se_mistral_l_n100_v1measured
2026-10-18se_mistral_m_n100_v1measured
2026-10-18se_mistral_s_n100_v1measured
2026-10-18se_openai_l_n100_v1measured
2026-10-18se_openai_m_n100_v1measured
2026-10-18se_openai_s_n100_v1measured
2026-10-18se_qwen37max_n100_v1measured
2026-10-18se_qwen_m_n100_v1measured
2026-10-18se_qwen_s_n100_v1measured
2026-10-18se_unieval_n100_v1measured
2026-10-18se_unievalpub_n100_v1measured
2026-10-15cnn_bart_s_dm200_v1measured
2026-10-15se_bartscore_n100_v1measured
2026-10-15se_coverage_n100_v1measured
2026-10-15se_length_n100_v1measured
2026-10-15se_random_n100_v1measured
2026-10-15se_rouge1_n100_v1measured
2026-10-15se_rouge2_n100_v1measured
2026-10-15se_rougeL_n100_v1measured
2026-10-14cnn_bart_l_dm200_v1measured
2026-10-14cnn_bart_l_xsum_dm200_v1measured
2026-10-14cnn_bart_m_dm200_v1measured
2026-10-14cnn_lead3_dm200_v1measured
2026-10-01fn_mistral_l_n200_v1rescored against gold/few_nerd_280
2026-09-30ag_bert_base_fy_n200_v1measured
2026-09-30ag_bert_base_ta_n200_v1measured
2026-09-303 arms on ag_news_200run again (linux)
2026-09-305 arms on cnn_dailymail_200run again (linux)
2026-09-30cnn_bart_l_xsum_n200_v1measured
2026-09-30cnn_bart_m_n200_v1measured
2026-09-30cnn_bart_s_n200_v1measured
2026-09-30cnn_bart_s_n200_v1 · repeatsrun again (repeats)
2026-09-30db_bert_base_fy_n200_v1measured
2026-09-30db_bert_base_fy_n200_v1 · linuxrun again (linux)
2026-09-30db_bert_base_fy_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30cnn_bart_l_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30cnn_bart_l_xsum_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30cnn_bart_m_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30cnn_bart_s_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30cnn_lead3_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30cnn_bart_s_n200_v1retired in r3: a second run of the record's config; data/runs is the record
2026-09-30ag_bert_base_fy_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30ag_bert_base_ta_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-30ag_bert_mini_n200_v1retired in r2: a re-timing of the record's config on one Linux machine; data/runs is the record
2026-09-293 arms on scifact_200re-parsed
2026-09-29sf_llama_l_n200_v1retired in r3: a second run of the record's config; data/runs is the record
2026-09-29sf_llama_s_n200_v1retired in r3: a second run of the record's config; data/runs is the record
2026-09-29sf_qwen_s_n200_v1retired in r3: a second run of the record's config; data/runs is the record
2026-09-2827 arms on few_nerd_280rescored against gold/few_nerd_280
2026-09-2823 arms on scifact_200re-parsed
2026-09-28sf_glm_s_bm25first_n200_v1 · pairretired in r2: a paired side run, not on the leaderboard; data/runs is the record
2026-09-28sf_glm_s_e5first_n200_v1 · pairretired in r2: a paired side run, not on the leaderboard; data/runs is the record
2026-09-28sf_qwen_s_bm25first_n200_v1 · pairretired in r2: a paired side run, not on the leaderboard; data/runs is the record
2026-09-28sf_qwen_s_e5first_n200_v1 · pairretired in r2: a paired side run, not on the leaderboard; data/runs is the record
2026-09-27ag_anthropic_l_n200_v1measured
2026-09-27ag_anthropic_m_n200_v1measured
2026-09-27ag_anthropic_s_n200_v1measured
2026-09-27ag_bart_mnli_n200_v1measured
2026-09-27ag_bert_mini_n200_v1measured
2026-09-27ag_constant_n200_v1measured
2026-09-27ag_deepseek_l_n200_v1measured
2026-09-27ag_deepseek_m_n200_v1measured
2026-09-27ag_deepseek_s_n200_v1measured
2026-09-27ag_gemma_l_n200_v1measured
2026-09-27ag_gemma_m_n200_v1measured
2026-09-27ag_gemma_s_n200_v1measured
2026-09-27ag_glm_l_n200_v1measured
2026-09-27ag_glm_m_n200_v1measured
2026-09-27ag_glm_s_n200_v1measured
2026-09-27ag_keyword_n200_v1measured
2026-09-27ag_llama_l_n200_v1measured
2026-09-27ag_llama_m_n200_v1measured
2026-09-27ag_llama_s_n200_v1measured
2026-09-27ag_mistral_l_n200_v1measured
2026-09-27ag_mistral_m_n200_v1measured
2026-09-27ag_mistral_s_n200_v1measured
2026-09-27ag_openai_l_n200_v1measured
2026-09-27ag_openai_m_n200_v1measured
2026-09-27ag_openai_s_n200_v1measured
2026-09-27ag_qwen_l_n200_v1measured
2026-09-27ag_qwen_m_n200_v1measured
2026-09-27ag_qwen_s_n200_v1measured
2026-09-27db_anthropic_l_n200_v1measured
2026-09-27db_anthropic_m_n200_v1measured
2026-09-27db_anthropic_s_n200_v1measured
2026-09-27db_bart_mnli_n200_v1measured
2026-09-27db_constant_n200_v1measured
2026-09-27db_deepseek_l_n200_v1measured
2026-09-27db_deepseek_m_n200_v1measured
2026-09-27db_deepseek_s_n200_v1measured
2026-09-27db_gemma_l_n200_v1measured
2026-09-27db_gemma_m_n200_v1measured
2026-09-27db_gemma_s_n200_v1measured
2026-09-27db_glm_l_n200_v1measured
2026-09-27db_glm_m_n200_v1measured
2026-09-27db_glm_s_n200_v1measured
2026-09-27db_keyword_n200_v1measured
2026-09-27db_llama_l_n200_v1measured
2026-09-27db_llama_m_n200_v1measured
2026-09-27db_llama_s_n200_v1measured
2026-09-27db_mistral_l_n200_v1measured
2026-09-27db_mistral_m_n200_v1measured
2026-09-27db_mistral_s_n200_v1measured
2026-09-27db_openai_l_n200_v1measured
2026-09-27db_openai_m_n200_v1measured
2026-09-27db_openai_s_n200_v1measured
2026-09-27db_qwen_l_n200_v1measured
2026-09-27db_qwen_m_n200_v1measured
2026-09-27db_qwen_s_n200_v1measured
2026-09-26cnn_bart_l_v1measured
2026-09-26cnn_lead3_v1measured
2026-09-26cnn_anthropic_l_n200_v1measured
2026-09-26cnn_anthropic_m_n200_v1measured
2026-09-26cnn_anthropic_s_n200_v1measured
2026-09-26cnn_bart_l_n200_v1measured
2026-09-26cnn_gemma_l_n200_v1measured
2026-09-26cnn_gemma_m_n200_v1measured
2026-09-26cnn_gemma_s_n200_v1measured
2026-09-26cnn_glm_l_n200_v1measured
2026-09-26cnn_glm_m_n200_v1measured
2026-09-26cnn_glm_s_n200_v1measured
2026-09-26cnn_lead3_n200_v1measured
2026-09-26cnn_llama_l_n200_v1measured
2026-09-26cnn_llama_m_n200_v1measured
2026-09-26cnn_llama_s_n200_v1measured
2026-09-26cnn_mistral_l_n200_v1measured
2026-09-26cnn_mistral_m_n200_v1measured
2026-09-26cnn_mistral_s_n200_v1measured
2026-09-26cnn_openai_l_n200_v1measured
2026-09-26cnn_openai_m_n200_v1measured
2026-09-26cnn_openai_s_n200_v1measured
2026-09-26cnn_qwen_l_n200_v1measured
2026-09-26cnn_qwen_s_n200_v1measured
2026-09-25cnn_deepseek_l_n200_v1measured
2026-09-25cnn_deepseek_m_n200_v1measured
2026-09-25cnn_deepseek_s_n200_v1measured
2026-09-25cnn_qwen_m_n200_v1measured

04cost and quality, by goal

$0.0000$0.0025cost (USD per item)0.7670.125fn_span_marker_n200_v1: f1 0.767, cost 0.000fn_openai_m_n200_v1: f1 0.686, cost 0.002fn_anthropic_l_n200_v1: f1 0.680, cost 0.002fn_gemma_m_n200_v1: f1 0.673, cost 0.000fn_openai_l_n200_v1: f1 0.672, cost 0.001fn_anthropic_m_n200_v1: f1 0.666, cost 0.001fn_qwen_s_n200_v1: f1 0.665, cost 0.000fn_qwen_m_n200_v1: f1 0.655, cost 0.000fn_gemma_l_n200_v1: f1 0.648, cost 0.000fn_qwen_l_n200_v1: f1 0.624, cost 0.000fn_deepseek_m_n200_v1: f1 0.615, cost 0.000fn_llama_l_n200_v1: f1 0.599, cost 0.000fn_mistral_l_n200_v1: f1 0.598, cost 0.000fn_openai_s_n200_v1: f1 0.595, cost 0.000fn_glm_m_n200_v1: f1 0.594, cost 0.000fn_mistral_m_n200_v1: f1 0.591, cost 0.000fn_deepseek_l_n200_v1: f1 0.587, cost 0.000fn_anthropic_s_n200_v1: f1 0.584, cost 0.001fn_deepseek_s_n200_v1: f1 0.576, cost 0.000fn_llama_m_n200_v1: f1 0.574, cost 0.000fn_llama_s_n200_v1: f1 0.569, cost 0.000fn_glm_l_n200_v1: f1 0.566, cost 0.000fn_glm_s_n200_v1: f1 0.553, cost 0.000fn_gemma_s_n200_v1: f1 0.550, cost 0.000fn_mistral_s_n200_v1: f1 0.524, cost 0.000fn_gliner_n200_v1: f1 0.454, cost 0.000fn_capitalized_n200_v1: f1 0.191, cost 0.000fn_nothing_n200_v1: f1 0.125, cost 0.000fn_span_marker_n200_v1
f1 against cost (USD per item): better is left top. Named and joined: the frontier, the arms no other arm beats on both.
0.00457450.5speed p95 (ms)0.7670.125fn_span_marker_n200_v1: f1 0.767, speed 1216.7fn_openai_m_n200_v1: f1 0.686, speed 2321.6fn_anthropic_l_n200_v1: f1 0.680, speed 4020.2fn_gemma_m_n200_v1: f1 0.673, speed 10254.5fn_openai_l_n200_v1: f1 0.672, speed 2890.1fn_anthropic_m_n200_v1: f1 0.666, speed 3392.4fn_qwen_s_n200_v1: f1 0.665, speed 3746.6fn_qwen_m_n200_v1: f1 0.655, speed 4116.2fn_gemma_l_n200_v1: f1 0.648, speed 7003.0fn_qwen_l_n200_v1: f1 0.624, speed 2670.8fn_deepseek_m_n200_v1: f1 0.615, speed 3161.3fn_llama_l_n200_v1: f1 0.599, speed 12665.1fn_mistral_l_n200_v1: f1 0.598, speed 57450.5fn_openai_s_n200_v1: f1 0.595, speed 1571.4fn_glm_m_n200_v1: f1 0.594, speed 7802.9fn_mistral_m_n200_v1: f1 0.591, speed 1293.9fn_deepseek_l_n200_v1: f1 0.587, speed 7751.9fn_anthropic_s_n200_v1: f1 0.584, speed 1901.8fn_deepseek_s_n200_v1: f1 0.576, speed 5276.7fn_llama_m_n200_v1: f1 0.574, speed 15183.2fn_llama_s_n200_v1: f1 0.569, speed 5652.6fn_glm_l_n200_v1: f1 0.566, speed 8248.7fn_glm_s_n200_v1: f1 0.553, speed 3881.7fn_gemma_s_n200_v1: f1 0.550, speed 13451.2fn_mistral_s_n200_v1: f1 0.524, speed 4100.1fn_gliner_n200_v1: f1 0.454, speed 238.7fn_capitalized_n200_v1: f1 0.191, speed 0.066fn_nothing_n200_v1: f1 0.125, speed 0.004fn_span_marker_n200_v1fn_gliner_n200_v1fn_capitalized_n200_v1fn_nothing_n200_v1
f1 against speed p95 (ms): better is left top. Named and joined: the frontier, the arms no other arm beats on both.