eval-harness

01what it measures, and how

Wall-clock time of one call, in milliseconds.

The wall-clock time of one call. When an adapter reports none, the core times the call itself; a model's first call, which loads it (the warm-up), is excluded. It depends on the machine, the provider's load and the network: compare it between arms measured under the same conditions, which the inspector's conditions line checks. Lower is better, and it needs no reference.

02on ag_news_200

gold/ag_news_200ag_constant_n200_v1: 0.0021ag_keyword_n200_v1: 0.053ag_bert_mini_n200_v1: 7.11ag_bert_mini_n200_v1 · linux: 20.9ag_bert_base_ta_n200_v1 · linux: 56.6ag_bert_base_ta_n200_v1: 76.3ag_bert_base_fy_n200_v1: 90.0ag_bert_base_fy_n200_v1 · linux: 201.5ag_mistral_m_n200_v1: 436.6ag_gemma_s_n200_v1: 664.2ag_gemma_m_n200_v1: 666.2ag_mistral_s_n200_v1: 731.9ag_llama_s_n200_v1: 784.0ag_llama_l_n200_v1: 804.5ag_openai_s_n200_v1: 859.0ag_deepseek_m_n200_v1: 889.7ag_qwen_m_n200_v1: 910.8ag_gemma_l_n200_v1: 911.1ag_anthropic_s_n200_v1: 987.0ag_openai_m_n200_v1: 1105.4ag_qwen_s_n200_v1: 1144.1ag_glm_s_n200_v1: 1153.8ag_llama_m_n200_v1: 1193.4ag_openai_l_n200_v1: 1207.3ag_deepseek_l_n200_v1: 1323.2ag_bart_mnli_n200_v1: 1386.5ag_deepseek_s_n200_v1: 1602.5ag_glm_l_n200_v1: 1870.9ag_anthropic_m_n200_v1: 1915.9ag_anthropic_l_n200_v1: 1965.1ag_glm_m_n200_v1: 2261.9ag_qwen_l_n200_v1: 2849.4ag_mistral_l_n200_v1: 7662.1gold/ag_news_200_adj1ag_constant_n200_v1 · adjudicated: 0.0021ag_keyword_n200_v1 · adjudicated: 0.053ag_bert_mini_n200_v1 · adjudicated: 7.11ag_bert_base_ta_n200_v1 · adjudicated: 76.3ag_bert_base_fy_n200_v1 · adjudicated: 90.0ag_mistral_m_n200_v1 · adjudicated: 436.6ag_gemma_s_n200_v1 · adjudicated: 664.2ag_gemma_m_n200_v1 · adjudicated: 666.2ag_mistral_s_n200_v1 · adjudicated: 731.9ag_llama_s_n200_v1 · adjudicated: 784.0ag_llama_l_n200_v1 · adjudicated: 804.5ag_openai_s_n200_v1 · adjudicated: 859.0ag_deepseek_m_n200_v1 · adjudicated: 889.7ag_qwen_m_n200_v1 · adjudicated: 910.8ag_gemma_l_n200_v1 · adjudicated: 911.1ag_anthropic_s_n200_v1 · adjudicated: 987.0ag_openai_m_n200_v1 · adjudicated: 1105.4ag_qwen_s_n200_v1 · adjudicated: 1144.1ag_glm_s_n200_v1 · adjudicated: 1153.8ag_llama_m_n200_v1 · adjudicated: 1193.4ag_openai_l_n200_v1 · adjudicated: 1207.3ag_deepseek_l_n200_v1 · adjudicated: 1323.2ag_bart_mnli_n200_v1 · adjudicated: 1386.5ag_deepseek_s_n200_v1 · adjudicated: 1602.5ag_glm_l_n200_v1 · adjudicated: 1870.9ag_anthropic_m_n200_v1 · adjudicated: 1915.9ag_anthropic_l_n200_v1 · adjudicated: 1965.1ag_glm_m_n200_v1 · adjudicated: 2261.9ag_qwen_l_n200_v1 · adjudicated: 2849.4ag_mistral_l_n200_v1 · adjudicated: 7662.10.0021 · better7662.1

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean latency_msitemsscored against
ag_constant_n200_v10.002200gold/ag_news_200 gold
ag_keyword_n200_v10.053200gold/ag_news_200 gold
ag_bert_mini_n200_v17.110200gold/ag_news_200 gold
ag_bert_mini_n200_v1 · linux20.9200gold/ag_news_200 gold
ag_bert_base_ta_n200_v1 · linux56.6200gold/ag_news_200 gold
ag_bert_base_ta_n200_v176.3200gold/ag_news_200 gold
ag_bert_base_fy_n200_v190.0200gold/ag_news_200 gold
ag_bert_base_fy_n200_v1 · linux201.5200gold/ag_news_200 gold
ag_mistral_m_n200_v1436.6200gold/ag_news_200 gold
ag_gemma_s_n200_v1664.2200gold/ag_news_200 gold
ag_gemma_m_n200_v1666.2200gold/ag_news_200 gold
ag_mistral_s_n200_v1731.9200gold/ag_news_200 gold
ag_llama_s_n200_v1784.0200gold/ag_news_200 gold
ag_llama_l_n200_v1804.5200gold/ag_news_200 gold
ag_openai_s_n200_v1859.0200gold/ag_news_200 gold
ag_deepseek_m_n200_v1889.7200gold/ag_news_200 gold
ag_qwen_m_n200_v1910.8200gold/ag_news_200 gold
ag_gemma_l_n200_v1911.1200gold/ag_news_200 gold
ag_anthropic_s_n200_v1987.0200gold/ag_news_200 gold
ag_openai_m_n200_v11105.4200gold/ag_news_200 gold
ag_qwen_s_n200_v11144.1200gold/ag_news_200 gold
ag_glm_s_n200_v11153.8200gold/ag_news_200 gold
ag_llama_m_n200_v11193.4200gold/ag_news_200 gold
ag_openai_l_n200_v11207.3200gold/ag_news_200 gold
ag_deepseek_l_n200_v11323.2200gold/ag_news_200 gold
ag_bart_mnli_n200_v11386.5200gold/ag_news_200 gold
ag_deepseek_s_n200_v11602.5200gold/ag_news_200 gold
ag_glm_l_n200_v11870.9200gold/ag_news_200 gold
ag_anthropic_m_n200_v11915.9200gold/ag_news_200 gold
ag_anthropic_l_n200_v11965.1200gold/ag_news_200 gold
ag_glm_m_n200_v12261.9200gold/ag_news_200 gold
ag_qwen_l_n200_v12849.4200gold/ag_news_200 gold
ag_mistral_l_n200_v17662.1200gold/ag_news_200 gold
ag_constant_n200_v1 · adjudicated0.002200gold/ag_news_200_adj1 gold
ag_keyword_n200_v1 · adjudicated0.053200gold/ag_news_200_adj1 gold
ag_bert_mini_n200_v1 · adjudicated7.110200gold/ag_news_200_adj1 gold
ag_bert_base_ta_n200_v1 · adjudicated76.3200gold/ag_news_200_adj1 gold
ag_bert_base_fy_n200_v1 · adjudicated90.0200gold/ag_news_200_adj1 gold
ag_mistral_m_n200_v1 · adjudicated436.6200gold/ag_news_200_adj1 gold
ag_gemma_s_n200_v1 · adjudicated664.2200gold/ag_news_200_adj1 gold
ag_gemma_m_n200_v1 · adjudicated666.2200gold/ag_news_200_adj1 gold
ag_mistral_s_n200_v1 · adjudicated731.9200gold/ag_news_200_adj1 gold
ag_llama_s_n200_v1 · adjudicated784.0200gold/ag_news_200_adj1 gold
ag_llama_l_n200_v1 · adjudicated804.5200gold/ag_news_200_adj1 gold
ag_openai_s_n200_v1 · adjudicated859.0200gold/ag_news_200_adj1 gold
ag_deepseek_m_n200_v1 · adjudicated889.7200gold/ag_news_200_adj1 gold
ag_qwen_m_n200_v1 · adjudicated910.8200gold/ag_news_200_adj1 gold
ag_gemma_l_n200_v1 · adjudicated911.1200gold/ag_news_200_adj1 gold
ag_anthropic_s_n200_v1 · adjudicated987.0200gold/ag_news_200_adj1 gold
ag_openai_m_n200_v1 · adjudicated1105.4200gold/ag_news_200_adj1 gold
ag_qwen_s_n200_v1 · adjudicated1144.1200gold/ag_news_200_adj1 gold
ag_glm_s_n200_v1 · adjudicated1153.8200gold/ag_news_200_adj1 gold
ag_llama_m_n200_v1 · adjudicated1193.4200gold/ag_news_200_adj1 gold
ag_openai_l_n200_v1 · adjudicated1207.3200gold/ag_news_200_adj1 gold
ag_deepseek_l_n200_v1 · adjudicated1323.2200gold/ag_news_200_adj1 gold
ag_bart_mnli_n200_v1 · adjudicated1386.5200gold/ag_news_200_adj1 gold
ag_deepseek_s_n200_v1 · adjudicated1602.5200gold/ag_news_200_adj1 gold
ag_glm_l_n200_v1 · adjudicated1870.9200gold/ag_news_200_adj1 gold
ag_anthropic_m_n200_v1 · adjudicated1915.9200gold/ag_news_200_adj1 gold
ag_anthropic_l_n200_v1 · adjudicated1965.1200gold/ag_news_200_adj1 gold
ag_glm_m_n200_v1 · adjudicated2261.9200gold/ag_news_200_adj1 gold
ag_qwen_l_n200_v1 · adjudicated2849.4200gold/ag_news_200_adj1 gold
ag_mistral_l_n200_v1 · adjudicated7662.1200gold/ag_news_200_adj1 gold

03on cnn_dailymail_20

gold/cnn_dailymail_20cnn_lead3_v1: 0.071cnn_bart_l_v1: 11861.30.071 · better11861.3

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean latency_msitemsscored against
cnn_lead3_v10.07120gold/cnn_dailymail_20 gold
cnn_bart_l_v111861.320gold/cnn_dailymail_20 gold

04on cnn_dailymail_200

gold/cnn_dailymail_200cnn_lead3_n200_v1: 0.065cnn_lead3_n200_v1 · linux: 0.07cnn_openai_s_n200_v1: 1276.6cnn_qwen_m_n200_v1: 1534.0cnn_deepseek_m_n200_v1: 1595.4cnn_glm_s_n200_v1: 1610.4cnn_openai_m_n200_v1: 1744.8cnn_openai_l_n200_v1: 1774.8cnn_gemma_m_n200_v1: 1872.5cnn_anthropic_s_n200_v1: 2025.0cnn_llama_s_n200_v1: 2388.0cnn_gemma_l_n200_v1: 2591.1cnn_mistral_m_n200_v1: 2653.7cnn_deepseek_l_n200_v1: 2665.5cnn_deepseek_s_n200_v1: 2752.6cnn_mistral_s_n200_v1: 2819.9cnn_qwen_l_n200_v1: 2944.9cnn_anthropic_m_n200_v1: 3034.2cnn_glm_m_n200_v1: 3230.1cnn_anthropic_l_n200_v1: 3402.0cnn_bart_s_n200_v1 · linux: 3420.4cnn_llama_l_n200_v1: 3473.0cnn_bart_s_n200_v1 · repeats: 3765.5cnn_llama_m_n200_v1: 4045.2cnn_bart_l_xsum_n200_v1 · linux: 4248.2cnn_bart_m_n200_v1 · linux: 4374.7cnn_bart_s_n200_v1: 4484.2cnn_bart_l_xsum_n200_v1: 4620.6cnn_bart_m_n200_v1: 5061.7cnn_glm_l_n200_v1: 5689.4cnn_bart_l_n200_v1 · linux: 6507.4cnn_gemma_s_n200_v1: 6956.2cnn_bart_l_n200_v1: 11009.6cnn_mistral_l_n200_v1: 11044.7cnn_qwen_s_n200_v1: 25037.7silver/cnn_dailymail_200cnn_lead3_n200_v1 · silver: 0.065cnn_openai_s_n200_v1 · silver: 1276.6cnn_qwen_m_n200_v1 · silver: 1534.0cnn_deepseek_m_n200_v1 · silver: 1595.4cnn_glm_s_n200_v1 · silver: 1610.4cnn_openai_m_n200_v1 · silver: 1744.8cnn_openai_l_n200_v1 · silver: 1774.8cnn_gemma_m_n200_v1 · silver: 1872.5cnn_anthropic_s_n200_v1 · silver: 2025.0cnn_llama_s_n200_v1 · silver: 2388.0cnn_gemma_l_n200_v1 · silver: 2591.1cnn_mistral_m_n200_v1 · silver: 2653.7cnn_deepseek_l_n200_v1 · silver: 2665.5cnn_deepseek_s_n200_v1 · silver: 2752.6cnn_mistral_s_n200_v1 · silver: 2819.9cnn_qwen_l_n200_v1 · silver: 2944.9cnn_anthropic_m_n200_v1 · silver: 3034.2cnn_glm_m_n200_v1 · silver: 3230.1cnn_llama_l_n200_v1 · silver: 3473.0cnn_llama_m_n200_v1 · silver: 4045.2cnn_bart_s_n200_v1 · silver: 4484.2cnn_bart_l_xsum_n200_v1 · silver: 4620.6cnn_bart_m_n200_v1 · silver: 5061.7cnn_glm_l_n200_v1 · silver: 5689.4cnn_gemma_s_n200_v1 · silver: 6956.2cnn_bart_l_n200_v1 · silver: 11009.6cnn_mistral_l_n200_v1 · silver: 11044.7cnn_qwen_s_n200_v1 · silver: 25037.7judge judge-summ-pick-v1cnn_lead3_n200_v1 · judged/judge-summ-pick-v1: 0.065cnn_openai_s_n200_v1 · judged/judge-summ-pick-v1: 1276.8cnn_qwen_m_n200_v1 · judged/judge-summ-pick-v1: 1533.2cnn_deepseek_m_n200_v1 · judged/judge-summ-pick-v1: 1592.0cnn_glm_s_n200_v1 · judged/judge-summ-pick-v1: 1611.8cnn_openai_m_n200_v1 · judged/judge-summ-pick-v1: 1745.7cnn_openai_l_n200_v1 · judged/judge-summ-pick-v1: 1774.7cnn_gemma_m_n200_v1 · judged/judge-summ-pick-v1: 1871.0cnn_anthropic_s_n200_v1 · judged/judge-summ-pick-v1: 2022.6cnn_llama_s_n200_v1 · judged/judge-summ-pick-v1: 2369.9cnn_gemma_l_n200_v1 · judged/judge-summ-pick-v1: 2600.4cnn_mistral_m_n200_v1 · judged/judge-summ-pick-v1: 2648.0cnn_deepseek_l_n200_v1 · judged/judge-summ-pick-v1: 2658.8cnn_deepseek_s_n200_v1 · judged/judge-summ-pick-v1: 2713.6cnn_mistral_s_n200_v1 · judged/judge-summ-pick-v1: 2810.1cnn_qwen_l_n200_v1 · judged/judge-summ-pick-v1: 2943.1cnn_anthropic_m_n200_v1 · judged/judge-summ-pick-v1: 3032.8cnn_glm_m_n200_v1 · judged/judge-summ-pick-v1: 3243.3cnn_anthropic_l_n200_v1 · judged/judge-summ-pick-v1: 3398.7cnn_llama_l_n200_v1 · judged/judge-summ-pick-v1: 3474.2cnn_llama_m_n200_v1 · judged/judge-summ-pick-v1: 4066.1cnn_bart_s_n200_v1 · judged/judge-summ-pick-v1: 4484.8cnn_bart_l_xsum_n200_v1 · judged/judge-summ-pick-v1: 4609.2cnn_bart_m_n200_v1 · judged/judge-summ-pick-v1: 5047.9cnn_glm_l_n200_v1 · judged/judge-summ-pick-v1: 5718.2cnn_gemma_s_n200_v1 · judged/judge-summ-pick-v1: 6932.0cnn_mistral_l_n200_v1 · judged/judge-summ-pick-v1: 10528.6cnn_bart_l_n200_v1 · judged/judge-summ-pick-v1: 11006.8cnn_qwen_s_n200_v1 · judged/judge-summ-pick-v1: 25217.6judge judge-summ-v1cnn_lead3_n200_v1 · judged/judge-summ-v1: 0.065cnn_openai_s_n200_v1 · judged/judge-summ-v1: 1276.8cnn_qwen_m_n200_v1 · judged/judge-summ-v1: 1531.7cnn_deepseek_m_n200_v1 · judged/judge-summ-v1: 1588.5cnn_glm_s_n200_v1 · judged/judge-summ-v1: 1615.3cnn_openai_m_n200_v1 · judged/judge-summ-v1: 1746.2cnn_openai_l_n200_v1 · judged/judge-summ-v1: 1769.2cnn_gemma_m_n200_v1 · judged/judge-summ-v1: 1873.0cnn_anthropic_s_n200_v1 · judged/judge-summ-v1: 2019.8cnn_llama_s_n200_v1 · judged/judge-summ-v1: 2391.7cnn_gemma_l_n200_v1 · judged/judge-summ-v1: 2601.4cnn_mistral_m_n200_v1 · judged/judge-summ-v1: 2658.6cnn_deepseek_l_n200_v1 · judged/judge-summ-v1: 2665.9cnn_deepseek_s_n200_v1 · judged/judge-summ-v1: 2752.8cnn_mistral_s_n200_v1 · judged/judge-summ-v1: 2826.3cnn_qwen_l_n200_v1 · judged/judge-summ-v1: 2946.3cnn_anthropic_m_n200_v1 · judged/judge-summ-v1: 3031.5cnn_glm_m_n200_v1 · judged/judge-summ-v1: 3234.1cnn_anthropic_l_n200_v1 · judged/judge-summ-v1: 3402.2cnn_llama_l_n200_v1 · judged/judge-summ-v1: 3495.5cnn_llama_m_n200_v1 · judged/judge-summ-v1: 4072.3cnn_bart_s_n200_v1 · judged/judge-summ-v1: 4488.9cnn_bart_l_xsum_n200_v1 · judged/judge-summ-v1: 4604.6cnn_bart_m_n200_v1 · judged/judge-summ-v1: 5054.3cnn_glm_l_n200_v1 · judged/judge-summ-v1: 5738.7cnn_gemma_s_n200_v1 · judged/judge-summ-v1: 6988.6cnn_bart_l_n200_v1 · judged/judge-summ-v1: 10995.7cnn_mistral_l_n200_v1 · judged/judge-summ-v1: 11179.5cnn_qwen_s_n200_v1 · judged/judge-summ-v1: 24871.6judge judge-summ-v2cnn_lead3_n200_v1 · judged/judge-summ-v2: 0.065cnn_openai_s_n200_v1 · judged/judge-summ-v2: 1276.6cnn_qwen_m_n200_v1 · judged/judge-summ-v2: 1534.0cnn_deepseek_m_n200_v1 · judged/judge-summ-v2: 1595.4cnn_glm_s_n200_v1 · judged/judge-summ-v2: 1610.4cnn_openai_m_n200_v1 · judged/judge-summ-v2: 1744.8cnn_openai_l_n200_v1 · judged/judge-summ-v2: 1774.8cnn_gemma_m_n200_v1 · judged/judge-summ-v2: 1872.5cnn_anthropic_s_n200_v1 · judged/judge-summ-v2: 2025.0cnn_llama_s_n200_v1 · judged/judge-summ-v2: 2388.0cnn_gemma_l_n200_v1 · judged/judge-summ-v2: 2591.1cnn_mistral_m_n200_v1 · judged/judge-summ-v2: 2653.7cnn_deepseek_l_n200_v1 · judged/judge-summ-v2: 2665.5cnn_deepseek_s_n200_v1 · judged/judge-summ-v2: 2752.6cnn_mistral_s_n200_v1 · judged/judge-summ-v2: 2819.9cnn_qwen_l_n200_v1 · judged/judge-summ-v2: 2944.9cnn_anthropic_m_n200_v1 · judged/judge-summ-v2: 3034.2cnn_glm_m_n200_v1 · judged/judge-summ-v2: 3230.1cnn_anthropic_l_n200_v1 · judged/judge-summ-v2: 3402.0cnn_llama_l_n200_v1 · judged/judge-summ-v2: 3473.0cnn_llama_m_n200_v1 · judged/judge-summ-v2: 4045.2cnn_bart_s_n200_v1 · judged/judge-summ-v2: 4484.2cnn_bart_l_xsum_n200_v1 · judged/judge-summ-v2: 4620.6cnn_bart_m_n200_v1 · judged/judge-summ-v2: 5061.7cnn_glm_l_n200_v1 · judged/judge-summ-v2: 5689.4cnn_gemma_s_n200_v1 · judged/judge-summ-v2: 6956.2cnn_bart_l_n200_v1 · judged/judge-summ-v2: 11009.6cnn_mistral_l_n200_v1 · judged/judge-summ-v2: 11044.7cnn_qwen_s_n200_v1 · judged/judge-summ-v2: 25037.7judge judge-summ-pointwise-v1cnn_deepseek_m_n200_v1 · judged/judge-summ-pointwise-v1: 1595.4cnn_bart_l_n200_v1 · judged/judge-summ-pointwise-v1: 11009.60.065 · better25217.6

Not drawn: cnn_anthropic_l_n200_v1 · silver, scored against the reference it wrote (its row is in the table).

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean latency_msitemsscored against
cnn_lead3_n200_v10.065200gold/cnn_dailymail_200 gold
cnn_lead3_n200_v1 · linux0.070200gold/cnn_dailymail_200 gold
cnn_openai_s_n200_v11276.6200gold/cnn_dailymail_200 gold
cnn_qwen_m_n200_v11534.0200gold/cnn_dailymail_200 gold
cnn_deepseek_m_n200_v11595.4200gold/cnn_dailymail_200 gold
cnn_glm_s_n200_v11610.4200gold/cnn_dailymail_200 gold
cnn_openai_m_n200_v11744.8200gold/cnn_dailymail_200 gold
cnn_openai_l_n200_v11774.8200gold/cnn_dailymail_200 gold
cnn_gemma_m_n200_v11872.5200gold/cnn_dailymail_200 gold
cnn_anthropic_s_n200_v12025.0200gold/cnn_dailymail_200 gold
cnn_llama_s_n200_v12388.0200gold/cnn_dailymail_200 gold
cnn_gemma_l_n200_v12591.1200gold/cnn_dailymail_200 gold
cnn_mistral_m_n200_v12653.7200gold/cnn_dailymail_200 gold
cnn_deepseek_l_n200_v12665.5200gold/cnn_dailymail_200 gold
cnn_deepseek_s_n200_v12752.6200gold/cnn_dailymail_200 gold
cnn_mistral_s_n200_v12819.9200gold/cnn_dailymail_200 gold
cnn_qwen_l_n200_v12944.9200gold/cnn_dailymail_200 gold
cnn_anthropic_m_n200_v13034.2200gold/cnn_dailymail_200 gold
cnn_glm_m_n200_v13230.1200gold/cnn_dailymail_200 gold
cnn_anthropic_l_n200_v13402.0200gold/cnn_dailymail_200 gold
cnn_bart_s_n200_v1 · linux3420.4200gold/cnn_dailymail_200 gold
cnn_llama_l_n200_v13473.0200gold/cnn_dailymail_200 gold
cnn_bart_s_n200_v1 · repeats3765.5200gold/cnn_dailymail_200 gold
cnn_llama_m_n200_v14045.2200gold/cnn_dailymail_200 gold
cnn_bart_l_xsum_n200_v1 · linux4248.2200gold/cnn_dailymail_200 gold
cnn_bart_m_n200_v1 · linux4374.7200gold/cnn_dailymail_200 gold
cnn_bart_s_n200_v14484.2200gold/cnn_dailymail_200 gold
cnn_bart_l_xsum_n200_v14620.6200gold/cnn_dailymail_200 gold
cnn_bart_m_n200_v15061.7200gold/cnn_dailymail_200 gold
cnn_glm_l_n200_v15689.4200gold/cnn_dailymail_200 gold
cnn_bart_l_n200_v1 · linux6507.4200gold/cnn_dailymail_200 gold
cnn_gemma_s_n200_v16956.2200gold/cnn_dailymail_200 gold
cnn_bart_l_n200_v111009.6200gold/cnn_dailymail_200 gold
cnn_mistral_l_n200_v111044.7200gold/cnn_dailymail_200 gold
cnn_qwen_s_n200_v125037.7200gold/cnn_dailymail_200 gold
cnn_lead3_n200_v1 · silver0.065200silver/cnn_dailymail_200 silver
cnn_openai_s_n200_v1 · silver1276.6200silver/cnn_dailymail_200 silver
cnn_qwen_m_n200_v1 · silver1534.0200silver/cnn_dailymail_200 silver
cnn_deepseek_m_n200_v1 · silver1595.4200silver/cnn_dailymail_200 silver
cnn_glm_s_n200_v1 · silver1610.4200silver/cnn_dailymail_200 silver
cnn_openai_m_n200_v1 · silver1744.8200silver/cnn_dailymail_200 silver
cnn_openai_l_n200_v1 · silver1774.8200silver/cnn_dailymail_200 silver
cnn_gemma_m_n200_v1 · silver1872.5200silver/cnn_dailymail_200 silver
cnn_anthropic_s_n200_v1 · silver2025.0200silver/cnn_dailymail_200 silver
cnn_llama_s_n200_v1 · silver2388.0200silver/cnn_dailymail_200 silver
cnn_gemma_l_n200_v1 · silver2591.1200silver/cnn_dailymail_200 silver
cnn_mistral_m_n200_v1 · silver2653.7200silver/cnn_dailymail_200 silver
cnn_deepseek_l_n200_v1 · silver2665.5200silver/cnn_dailymail_200 silver
cnn_deepseek_s_n200_v1 · silver2752.6200silver/cnn_dailymail_200 silver
cnn_mistral_s_n200_v1 · silver2819.9200silver/cnn_dailymail_200 silver
cnn_qwen_l_n200_v1 · silver2944.9200silver/cnn_dailymail_200 silver
cnn_anthropic_m_n200_v1 · silver3034.2200silver/cnn_dailymail_200 silver
cnn_glm_m_n200_v1 · silver3230.1200silver/cnn_dailymail_200 silver
cnn_anthropic_l_n200_v1 · silver3402.0200silver/cnn_dailymail_200 silver
cnn_llama_l_n200_v1 · silver3473.0200silver/cnn_dailymail_200 silver
cnn_llama_m_n200_v1 · silver4045.2200silver/cnn_dailymail_200 silver
cnn_bart_s_n200_v1 · silver4484.2200silver/cnn_dailymail_200 silver
cnn_bart_l_xsum_n200_v1 · silver4620.6200silver/cnn_dailymail_200 silver
cnn_bart_m_n200_v1 · silver5061.7200silver/cnn_dailymail_200 silver
cnn_glm_l_n200_v1 · silver5689.4200silver/cnn_dailymail_200 silver
cnn_gemma_s_n200_v1 · silver6956.2200silver/cnn_dailymail_200 silver
cnn_bart_l_n200_v1 · silver11009.6200silver/cnn_dailymail_200 silver
cnn_mistral_l_n200_v1 · silver11044.7200silver/cnn_dailymail_200 silver
cnn_qwen_s_n200_v1 · silver25037.7200silver/cnn_dailymail_200 silver
cnn_lead3_n200_v1 · judged/judge-summ-pick-v10.065198judge judge-summ-pick-v1
cnn_lead3_n200_v1 · judged/judge-summ-v10.065197judge judge-summ-v1
cnn_lead3_n200_v1 · judged/judge-summ-v20.065200judge judge-summ-v2
cnn_openai_s_n200_v1 · judged/judge-summ-v21276.6200judge judge-summ-v2
cnn_openai_s_n200_v1 · judged/judge-summ-pick-v11276.8198judge judge-summ-pick-v1
cnn_openai_s_n200_v1 · judged/judge-summ-v11276.8197judge judge-summ-v1
cnn_qwen_m_n200_v1 · judged/judge-summ-v11531.7197judge judge-summ-v1
cnn_qwen_m_n200_v1 · judged/judge-summ-pick-v11533.2198judge judge-summ-pick-v1
cnn_qwen_m_n200_v1 · judged/judge-summ-v21534.0200judge judge-summ-v2
cnn_deepseek_m_n200_v1 · judged/judge-summ-v11588.5197judge judge-summ-v1
cnn_deepseek_m_n200_v1 · judged/judge-summ-pick-v11592.0198judge judge-summ-pick-v1
cnn_deepseek_m_n200_v1 · judged/judge-summ-pointwise-v11595.4200judge judge-summ-pointwise-v1
cnn_deepseek_m_n200_v1 · judged/judge-summ-v21595.4200judge judge-summ-v2
cnn_glm_s_n200_v1 · judged/judge-summ-v21610.4200judge judge-summ-v2
cnn_glm_s_n200_v1 · judged/judge-summ-pick-v11611.8198judge judge-summ-pick-v1
cnn_glm_s_n200_v1 · judged/judge-summ-v11615.3197judge judge-summ-v1
cnn_openai_m_n200_v1 · judged/judge-summ-v21744.8200judge judge-summ-v2
cnn_openai_m_n200_v1 · judged/judge-summ-pick-v11745.7198judge judge-summ-pick-v1
cnn_openai_m_n200_v1 · judged/judge-summ-v11746.2197judge judge-summ-v1
cnn_openai_l_n200_v1 · judged/judge-summ-v11769.2197judge judge-summ-v1
cnn_openai_l_n200_v1 · judged/judge-summ-pick-v11774.7198judge judge-summ-pick-v1
cnn_openai_l_n200_v1 · judged/judge-summ-v21774.8200judge judge-summ-v2
cnn_gemma_m_n200_v1 · judged/judge-summ-pick-v11871.0198judge judge-summ-pick-v1
cnn_gemma_m_n200_v1 · judged/judge-summ-v21872.5200judge judge-summ-v2
cnn_gemma_m_n200_v1 · judged/judge-summ-v11873.0197judge judge-summ-v1
cnn_anthropic_s_n200_v1 · judged/judge-summ-v12019.8197judge judge-summ-v1
cnn_anthropic_s_n200_v1 · judged/judge-summ-pick-v12022.6198judge judge-summ-pick-v1
cnn_anthropic_s_n200_v1 · judged/judge-summ-v22025.0200judge judge-summ-v2
cnn_llama_s_n200_v1 · judged/judge-summ-pick-v12369.9198judge judge-summ-pick-v1
cnn_llama_s_n200_v1 · judged/judge-summ-v22388.0200judge judge-summ-v2
cnn_llama_s_n200_v1 · judged/judge-summ-v12391.7197judge judge-summ-v1
cnn_gemma_l_n200_v1 · judged/judge-summ-v22591.1200judge judge-summ-v2
cnn_gemma_l_n200_v1 · judged/judge-summ-pick-v12600.4198judge judge-summ-pick-v1
cnn_gemma_l_n200_v1 · judged/judge-summ-v12601.4197judge judge-summ-v1
cnn_mistral_m_n200_v1 · judged/judge-summ-pick-v12648.0198judge judge-summ-pick-v1
cnn_mistral_m_n200_v1 · judged/judge-summ-v22653.7200judge judge-summ-v2
cnn_mistral_m_n200_v1 · judged/judge-summ-v12658.6197judge judge-summ-v1
cnn_deepseek_l_n200_v1 · judged/judge-summ-pick-v12658.8198judge judge-summ-pick-v1
cnn_deepseek_l_n200_v1 · judged/judge-summ-v22665.5200judge judge-summ-v2
cnn_deepseek_l_n200_v1 · judged/judge-summ-v12665.9197judge judge-summ-v1
cnn_deepseek_s_n200_v1 · judged/judge-summ-pick-v12713.6198judge judge-summ-pick-v1
cnn_deepseek_s_n200_v1 · judged/judge-summ-v22752.6200judge judge-summ-v2
cnn_deepseek_s_n200_v1 · judged/judge-summ-v12752.8197judge judge-summ-v1
cnn_mistral_s_n200_v1 · judged/judge-summ-pick-v12810.1198judge judge-summ-pick-v1
cnn_mistral_s_n200_v1 · judged/judge-summ-v22819.9200judge judge-summ-v2
cnn_mistral_s_n200_v1 · judged/judge-summ-v12826.3197judge judge-summ-v1
cnn_qwen_l_n200_v1 · judged/judge-summ-pick-v12943.1198judge judge-summ-pick-v1
cnn_qwen_l_n200_v1 · judged/judge-summ-v22944.9200judge judge-summ-v2
cnn_qwen_l_n200_v1 · judged/judge-summ-v12946.3197judge judge-summ-v1
cnn_anthropic_m_n200_v1 · judged/judge-summ-v13031.5197judge judge-summ-v1
cnn_anthropic_m_n200_v1 · judged/judge-summ-pick-v13032.8198judge judge-summ-pick-v1
cnn_anthropic_m_n200_v1 · judged/judge-summ-v23034.2200judge judge-summ-v2
cnn_glm_m_n200_v1 · judged/judge-summ-v23230.1200judge judge-summ-v2
cnn_glm_m_n200_v1 · judged/judge-summ-v13234.1197judge judge-summ-v1
cnn_glm_m_n200_v1 · judged/judge-summ-pick-v13243.3198judge judge-summ-pick-v1
cnn_anthropic_l_n200_v1 · judged/judge-summ-pick-v13398.7198judge judge-summ-pick-v1
cnn_anthropic_l_n200_v1 · judged/judge-summ-v23402.0200judge judge-summ-v2
cnn_anthropic_l_n200_v1 · judged/judge-summ-v13402.2197judge judge-summ-v1
cnn_llama_l_n200_v1 · judged/judge-summ-v23473.0200judge judge-summ-v2
cnn_llama_l_n200_v1 · judged/judge-summ-pick-v13474.2198judge judge-summ-pick-v1
cnn_llama_l_n200_v1 · judged/judge-summ-v13495.5197judge judge-summ-v1
cnn_llama_m_n200_v1 · judged/judge-summ-v24045.2200judge judge-summ-v2
cnn_llama_m_n200_v1 · judged/judge-summ-pick-v14066.1198judge judge-summ-pick-v1
cnn_llama_m_n200_v1 · judged/judge-summ-v14072.3197judge judge-summ-v1
cnn_bart_s_n200_v1 · judged/judge-summ-v24484.2200judge judge-summ-v2
cnn_bart_s_n200_v1 · judged/judge-summ-pick-v14484.8198judge judge-summ-pick-v1
cnn_bart_s_n200_v1 · judged/judge-summ-v14488.9197judge judge-summ-v1
cnn_bart_l_xsum_n200_v1 · judged/judge-summ-v14604.6197judge judge-summ-v1
cnn_bart_l_xsum_n200_v1 · judged/judge-summ-pick-v14609.2198judge judge-summ-pick-v1
cnn_bart_l_xsum_n200_v1 · judged/judge-summ-v24620.6200judge judge-summ-v2
cnn_bart_m_n200_v1 · judged/judge-summ-pick-v15047.9198judge judge-summ-pick-v1
cnn_bart_m_n200_v1 · judged/judge-summ-v15054.3197judge judge-summ-v1
cnn_bart_m_n200_v1 · judged/judge-summ-v25061.7200judge judge-summ-v2
cnn_glm_l_n200_v1 · judged/judge-summ-v25689.4200judge judge-summ-v2
cnn_glm_l_n200_v1 · judged/judge-summ-pick-v15718.2198judge judge-summ-pick-v1
cnn_glm_l_n200_v1 · judged/judge-summ-v15738.7197judge judge-summ-v1
cnn_gemma_s_n200_v1 · judged/judge-summ-pick-v16932.0198judge judge-summ-pick-v1
cnn_gemma_s_n200_v1 · judged/judge-summ-v26956.2200judge judge-summ-v2
cnn_gemma_s_n200_v1 · judged/judge-summ-v16988.6197judge judge-summ-v1
cnn_mistral_l_n200_v1 · judged/judge-summ-pick-v110528.6198judge judge-summ-pick-v1
cnn_bart_l_n200_v1 · judged/judge-summ-v110995.7197judge judge-summ-v1
cnn_bart_l_n200_v1 · judged/judge-summ-pick-v111006.8198judge judge-summ-pick-v1
cnn_bart_l_n200_v1 · judged/judge-summ-pointwise-v111009.6200judge judge-summ-pointwise-v1
cnn_bart_l_n200_v1 · judged/judge-summ-v211009.6200judge judge-summ-v2
cnn_mistral_l_n200_v1 · judged/judge-summ-v211044.7200judge judge-summ-v2
cnn_mistral_l_n200_v1 · judged/judge-summ-v111179.5197judge judge-summ-v1
cnn_qwen_s_n200_v1 · judged/judge-summ-v124871.6197judge judge-summ-v1
cnn_qwen_s_n200_v1 · judged/judge-summ-v225037.7200judge judge-summ-v2
cnn_qwen_s_n200_v1 · judged/judge-summ-pick-v125217.6198judge judge-summ-pick-v1

05on cnn_dailymail_dm200

gold/cnn_dailymail_dm200cnn_lead3_dm200_v1: 0.123cnn_bart_s_dm200_v1: 4302.1cnn_bart_l_xsum_dm200_v1: 5362.0cnn_bart_m_dm200_v1: 7292.4cnn_bart_l_dm200_v1: 11259.50.123 · better11259.5

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

06on dbpedia_280

gold/dbpedia_280db_constant_n200_v1: 0.0032db_keyword_n200_v1: 0.1db_bert_base_fy_n200_v1: 86.7db_bert_base_fy_n200_v1 · linux: 165.8db_mistral_m_n200_v1: 439.5db_gemma_s_n200_v1: 647.2db_llama_l_n200_v1: 659.8db_qwen_m_n200_v1: 853.2db_gemma_m_n200_v1: 856.5db_openai_s_n200_v1: 909.8db_mistral_s_n200_v1: 942.7db_deepseek_m_n200_v1: 974.6db_anthropic_s_n200_v1: 1082.6db_gemma_l_n200_v1: 1145.4db_openai_m_n200_v1: 1178.0db_glm_s_n200_v1: 1183.0db_openai_l_n200_v1: 1266.3db_llama_s_n200_v1: 1374.5db_deepseek_s_n200_v1: 1548.3db_qwen_l_n200_v1: 1660.0db_deepseek_l_n200_v1: 1781.5db_llama_m_n200_v1: 1927.0db_glm_m_n200_v1: 1930.1db_glm_l_n200_v1: 1978.0db_anthropic_m_n200_v1: 2257.9db_anthropic_l_n200_v1: 3101.1db_bart_mnli_n200_v1: 5717.7db_mistral_l_n200_v1: 10565.4db_qwen_s_n200_v1: 12349.10.0032 · better12349.1

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

07on few_nerd_280

gold/few_nerd_280fn_nothing_n200_v1: 0.0023fn_capitalized_n200_v1: 0.035fn_gliner_n200_v1: 169.4fn_span_marker_n200_v1: 757.9fn_mistral_m_n200_v1: 793.7fn_openai_s_n200_v1: 1205.6fn_deepseek_m_n200_v1: 1228.7fn_anthropic_s_n200_v1: 1337.5fn_openai_m_n200_v1: 1598.1fn_qwen_l_n200_v1: 1667.2fn_openai_l_n200_v1: 1903.2fn_mistral_s_n200_v1: 2063.6fn_glm_s_n200_v1: 2110.7fn_qwen_s_n200_v1: 2153.7fn_qwen_m_n200_v1: 2241.4fn_deepseek_s_n200_v1: 2277.1fn_llama_s_n200_v1: 2637.6fn_anthropic_m_n200_v1: 2691.7fn_anthropic_l_n200_v1: 2759.0fn_gemma_l_n200_v1: 2910.5fn_glm_l_n200_v1: 3013.7fn_gemma_m_n200_v1: 3049.6fn_glm_m_n200_v1: 3658.9fn_deepseek_l_n200_v1: 3726.8fn_llama_l_n200_v1: 4796.9fn_llama_m_n200_v1: 5402.1fn_gemma_s_n200_v1: 5941.2fn_mistral_l_n200_v1: 10247.10.0023 · better10247.1

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

08on fleet_dedup_v1

silver/fleet_dedup_v1dedup_guard_v1: 0.136no referencededup_pinned_f07efa9_v1: 0.1740.136 · better0.174

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean latency_msitemsscored against
dedup_guard_v10.136100silver/fleet_dedup_v1 silver
dedup_pinned_f07efa9_v10.174100—

09on orrery_triage_v1_labelled

gold/orrery_triage_v1_labelledtriage_stub_escalate_v1: 0.09triage_stub_escalate_v1 · rescored 3d3a6afd: 0.09triage_stub_escalate_v1 · rescored 5f2c057c: 0.09triage_stub_escalate_v1 · rescored 77adc97b: 0.09triage_stub_escalate_v1 · rescored cd35dc1f: 0.09triage_stub_dismiss_v1: 0.103triage_stub_dismiss_v1 · rescored 3d3a6afd: 0.103triage_stub_dismiss_v1 · rescored 5f2c057c: 0.103triage_stub_dismiss_v1 · rescored 77adc97b: 0.103triage_stub_dismiss_v1 · rescored cd35dc1f: 0.103triage_deepseek_41flash_v2: 3281.9triage_deepseek_41flash_v2 · rescored 3d3a6afd: 3281.9triage_deepseek_41flash_v2 · rescored 5f2c057c: 3281.9triage_gemini_35_flash_lite_v1: 3487.3triage_gemini_35_flash_lite_v1 · rescored 3d3a6afd: 3487.3triage_gemini_35_flash_lite_v1 · rescored 5f2c057c: 3487.3triage_gemini_35_flash_lite_v1 · rescored 77adc97b: 3487.3triage_gemini_35_flash_lite_v1 · rescored cd35dc1f: 3487.3triage_local_dgx_nemotron_35_bolt_e4b_v2: 3880.9triage_local_dgx_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afd: 3880.9triage_local_dgx_nemotron_35_bolt_e4b_v2 · rescored 5f2c057c: 3880.9triage_local_dgx_gpt_oss_20b_v2: 4680.7triage_local_dgx_gpt_oss_20b_v2 · rescored 3d3a6afd: 4680.7triage_local_dgx_gpt_oss_20b_v2 · rescored 5f2c057c: 4680.7triage_local_x3_nemotron_35_bolt_e4b_v2: 5646.0triage_local_x3_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afd: 5646.0triage_local_x3_nemotron_35_bolt_e4b_v2 · rescored 5f2c057c: 5646.0triage_local_dgx_gemma4_12b_noreason_v1: 5697.4triage_local_dgx_gemma4_12b_noreason_v1 · rescored 3d3a6afd: 5697.4triage_local_dgx_gemma4_12b_noreason_v1 · rescored 5f2c057c: 5697.4triage_local_dgx_gemma4_12b_noreason_v1 · rescored 77adc97b: 5697.4triage_local_dgx_gemma4_12b_noreason_v1 · rescored cd35dc1f: 5697.4triage_local_dgx_llama31_8b_v1: 5766.0triage_local_dgx_llama31_8b_v1 · rescored 3d3a6afd: 5766.0triage_local_dgx_llama31_8b_v1 · rescored 5f2c057c: 5766.0triage_local_dgx_llama31_8b_v1 · rescored 77adc97b: 5766.0triage_local_dgx_llama31_8b_v1 · rescored cd35dc1f: 5766.0triage_local_mac_gpt_oss_20b_v2: 5802.5triage_local_mac_gpt_oss_20b_v2 · rescored 3d3a6afd: 5802.5triage_local_mac_gpt_oss_20b_v2 · rescored 5f2c057c: 5802.5triage_local_x3_llama31_8b_v1: 5836.8triage_local_x3_llama31_8b_v1 · rescored 3d3a6afd: 5836.8triage_local_x3_llama31_8b_v1 · rescored 5f2c057c: 5836.8triage_local_x3_llama31_8b_v1 · rescored 77adc97b: 5836.8triage_local_x3_llama31_8b_v1 · rescored cd35dc1f: 5836.8triage_local_dgx_qwen3_30b_a3b_v1: 5859.1triage_local_dgx_qwen3_30b_a3b_v1 · rescored 3d3a6afd: 5859.1triage_local_dgx_qwen3_30b_a3b_v1 · rescored 5f2c057c: 5859.1triage_local_dgx_qwen3_30b_a3b_v1 · rescored 77adc97b: 5859.1triage_local_dgx_qwen3_30b_a3b_v1 · rescored cd35dc1f: 5859.1triage_local_x3_gpt_oss_20b_v2: 5863.5triage_local_x3_gpt_oss_20b_v2 · rescored 3d3a6afd: 5863.5triage_local_x3_gpt_oss_20b_v2 · rescored 5f2c057c: 5863.5triage_local_mac_nemotron_35_bolt_e4b_v2: 5883.1triage_local_mac_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afd: 5883.1triage_local_mac_nemotron_35_bolt_e4b_v2 · rescored 5f2c057c: 5883.1triage_hosted_mistral_medium_35_v1: 6103.3triage_hosted_mistral_medium_35_v1 · rescored 3d3a6afd: 6103.3triage_hosted_mistral_medium_35_v1 · rescored 5f2c057c: 6103.3triage_hosted_mistral_medium_35_v1 · rescored 77adc97b: 6103.3triage_hosted_mistral_medium_35_v1 · rescored cd35dc1f: 6103.3triage_hosted_gpt_oss_120b_v2: 6540.7triage_hosted_gpt_oss_120b_v2 · rescored 3d3a6afd: 6540.7triage_hosted_gpt_oss_120b_v2 · rescored 5f2c057c: 6540.7triage_hosted_qwen35_122b_v2: 6681.4triage_hosted_qwen35_122b_v2 · rescored 3d3a6afd: 6681.4triage_hosted_qwen35_122b_v2 · rescored 5f2c057c: 6681.4triage_local_dgx_nemotron_35_lightning_30b_v2: 6785.7triage_local_dgx_nemotron_35_lightning_30b_v2 · rescored 3d3a6afd: 6785.7triage_local_dgx_nemotron_35_lightning_30b_v2 · rescored 5f2c057c: 6785.7triage_local_x3_qwen3_30b_a3b_v1: 6932.3triage_local_x3_qwen3_30b_a3b_v1 · rescored 3d3a6afd: 6932.3triage_local_x3_qwen3_30b_a3b_v1 · rescored 5f2c057c: 6932.3triage_local_x3_qwen3_30b_a3b_v1 · rescored 77adc97b: 6932.3triage_local_x3_qwen3_30b_a3b_v1 · rescored cd35dc1f: 6932.3triage_local_dgx_gpt_oss_120b_v2: 7391.1triage_local_dgx_gpt_oss_120b_v2 · rescored 3d3a6afd: 7391.1triage_local_dgx_gpt_oss_120b_v2 · rescored 5f2c057c: 7391.1triage_haiku_55_v1 · repeats ×3: 7528.9triage_haiku_55_v1 · rescored 3d3a6afd, repeats ×3: 7528.9triage_haiku_55_v1 · rescored 5f2c057c, repeats ×3: 7528.9triage_haiku_55_v1 · rescored 77adc97b, repeats ×3: 7528.9triage_haiku_55_v1 · rescored cd35dc1f, repeats ×3: 7528.9triage_haiku_55_v1: 7625.9triage_haiku_55_v1 · rescored 3d3a6afd: 7625.9triage_haiku_55_v1 · rescored 5f2c057c: 7625.9triage_haiku_55_v1 · rescored 77adc97b: 7625.9triage_haiku_55_v1 · rescored cd35dc1f: 7625.9triage_local_mac_llama31_8b_v1: 7780.0triage_local_mac_llama31_8b_v1 · rescored 3d3a6afd: 7780.0triage_local_mac_llama31_8b_v1 · rescored 5f2c057c: 7780.0triage_local_mac_llama31_8b_v1 · rescored 77adc97b: 7780.0triage_local_mac_llama31_8b_v1 · rescored cd35dc1f: 7780.0triage_local_x3_nemotron_35_lightning_30b_v2: 7797.3triage_local_x3_nemotron_35_lightning_30b_v2 · rescored 3d3a6afd: 7797.3triage_local_x3_nemotron_35_lightning_30b_v2 · rescored 5f2c057c: 7797.3triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1: 7824.5triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 3d3a6afd: 7824.5triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 5f2c057c: 7824.5triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 77adc97b: 7824.5triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored cd35dc1f: 7824.5triage_gpt_6_luna_v2: 7973.3triage_gpt_6_luna_v2 · rescored 3d3a6afd: 7973.3triage_gpt_6_luna_v2 · rescored 5f2c057c: 7973.3triage_baseline_pro_noreason_v1: 8352.2triage_baseline_pro_noreason_v1 · rescored 3d3a6afd: 8352.2triage_baseline_pro_noreason_v1 · rescored 5f2c057c: 8352.2triage_local_mac_qwen3_30b_a3b_v1: 8996.0triage_local_mac_qwen3_30b_a3b_v1 · rescored 3d3a6afd: 8996.0triage_local_mac_qwen3_30b_a3b_v1 · rescored 5f2c057c: 8996.0triage_local_mac_qwen3_30b_a3b_v1 · rescored 77adc97b: 8996.0triage_local_mac_qwen3_30b_a3b_v1 · rescored cd35dc1f: 8996.0triage_local_mac_gemma4_12b_noreason_v1: 9174.2triage_local_mac_gemma4_12b_noreason_v1 · rescored 3d3a6afd: 9174.2triage_local_mac_gemma4_12b_noreason_v1 · rescored 5f2c057c: 9174.2triage_local_mac_gemma4_12b_noreason_v1 · rescored 77adc97b: 9174.2triage_local_mac_gemma4_12b_noreason_v1 · rescored cd35dc1f: 9174.2triage_deepseek_41flash_v1: 9832.8triage_deepseek_41flash_v1 · rescored 5f2c057c: 9832.8triage_deepseek_41flash_v1 · rescored 77adc97b: 9832.8triage_deepseek_41flash_v1 · rescored cd35dc1f: 9832.8triage_hosted_gemma4_31b_v1: 10510.3triage_hosted_gemma4_31b_v1 · rescored 5f2c057c: 10510.3triage_hosted_gemma4_31b_v1 · rescored 77adc97b: 10510.3triage_hosted_gemma4_31b_v1 · rescored cd35dc1f: 10510.3triage_hosted_gemma4_31b_v2: 10549.7triage_hosted_gemma4_31b_v2 · rescored 3d3a6afd: 10549.7triage_hosted_gemma4_31b_v2 · rescored 5f2c057c: 10549.7triage_local_mac_nemotron_35_lightning_30b_v2: 11064.9triage_local_mac_nemotron_35_lightning_30b_v2 · rescored 3d3a6afd: 11064.9triage_local_mac_nemotron_35_lightning_30b_v2 · rescored 5f2c057c: 11064.9triage_gpt_6_luna_v1 · repeats ×3: 11597.7triage_gpt_6_luna_v1 · rescored 5f2c057c, repeats ×3: 11597.7triage_gpt_6_luna_v1 · rescored 77adc97b, repeats ×3: 11597.7triage_gpt_6_luna_v1 · rescored cd35dc1f, repeats ×3: 11597.7triage_gpt_61_sol_v1 · repeats ×3: 12005.5triage_gpt_61_sol_v1 · rescored 3d3a6afd, repeats ×3: 12005.5triage_gpt_61_sol_v1 · rescored 5f2c057c, repeats ×3: 12005.5triage_gpt_61_sol_v1 · rescored 77adc97b, repeats ×3: 12005.5triage_gpt_61_sol_v1 · rescored cd35dc1f, repeats ×3: 12005.5triage_sonnet_55_v1: 12035.0triage_sonnet_55_v1 · rescored 3d3a6afd: 12035.0triage_sonnet_55_v1 · rescored 5f2c057c: 12035.0triage_sonnet_55_v1 · rescored 77adc97b: 12035.0triage_sonnet_55_v1 · rescored cd35dc1f: 12035.0triage_gpt_6_luna_v1: 12200.5triage_gpt_6_luna_v1 · rescored 5f2c057c: 12200.5triage_gpt_6_luna_v1 · rescored 77adc97b: 12200.5triage_gpt_6_luna_v1 · rescored cd35dc1f: 12200.5triage_gpt_61_sol_v1: 12241.4triage_gpt_61_sol_v1 · rescored 3d3a6afd: 12241.4triage_gpt_61_sol_v1 · rescored 5f2c057c: 12241.4triage_gpt_61_sol_v1 · rescored 77adc97b: 12241.4triage_gpt_61_sol_v1 · rescored cd35dc1f: 12241.4triage_local_dgx_gemma4_31b_v2: 13030.5triage_local_dgx_gemma4_31b_v2 · rescored 3d3a6afd: 13030.5triage_local_dgx_gemma4_31b_v2 · rescored 5f2c057c: 13030.5triage_local_dgx_llama4_scout_v1: 13354.0triage_local_dgx_llama4_scout_v1 · rescored 3d3a6afd: 13354.0triage_local_dgx_llama4_scout_v1 · rescored 5f2c057c: 13354.0triage_local_dgx_llama4_scout_v1 · rescored 77adc97b: 13354.0triage_local_dgx_llama4_scout_v1 · rescored cd35dc1f: 13354.0triage_local_dgx_gpt_oss_20b_v1: 13455.9triage_local_dgx_gpt_oss_20b_v1 · rescored 5f2c057c: 13455.9triage_local_dgx_gpt_oss_20b_v1 · rescored 77adc97b: 13455.9triage_local_dgx_gpt_oss_20b_v1 · rescored cd35dc1f: 13455.9triage_local_x3_gemma4_31b_v2: 14320.1triage_local_x3_gemma4_31b_v2 · rescored 3d3a6afd: 14320.1triage_local_x3_gemma4_31b_v2 · rescored 5f2c057c: 14320.1triage_local_dgx_qwen35_122b_v2: 14511.3triage_local_dgx_qwen35_122b_v2 · rescored 3d3a6afd: 14511.3triage_local_dgx_qwen35_122b_v2 · rescored 5f2c057c: 14511.3triage_local_dgx_gpt_oss_120b_v1: 14765.6triage_local_dgx_gpt_oss_120b_v1 · rescored 5f2c057c: 14765.6triage_local_dgx_gpt_oss_120b_v1 · rescored 77adc97b: 14765.6triage_local_dgx_gpt_oss_120b_v1 · rescored cd35dc1f: 14765.6triage_local_dgx_nemotron_35_bolt_e4b_v1: 15258.8triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c: 15258.8triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored 77adc97b: 15258.8triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f: 15258.8triage_local_dgx_nemotron_35_super_120b_v2: 15897.7triage_local_dgx_nemotron_35_super_120b_v2 · rescored 3d3a6afd: 15897.7triage_local_dgx_nemotron_35_super_120b_v2 · rescored 5f2c057c: 15897.7triage_local_dgx_qwen38_27b_v2: 16558.9triage_local_dgx_qwen38_27b_v2 · rescored 3d3a6afd: 16558.9triage_local_dgx_qwen38_27b_v2 · rescored 5f2c057c: 16558.9triage_local_x3_phi4_14b_v1: 16911.2triage_local_x3_phi4_14b_v1 · rescored 3d3a6afd: 16911.2triage_local_x3_phi4_14b_v1 · rescored 5f2c057c: 16911.2triage_local_x3_phi4_14b_v1 · rescored 77adc97b: 16911.2triage_local_x3_phi4_14b_v1 · rescored cd35dc1f: 16911.2triage_local_dgx_phi4_14b_v1: 16973.1triage_local_dgx_phi4_14b_v1 · rescored 3d3a6afd: 16973.1triage_local_dgx_phi4_14b_v1 · rescored 5f2c057c: 16973.1triage_local_dgx_phi4_14b_v1 · rescored 77adc97b: 16973.1triage_local_dgx_phi4_14b_v1 · rescored cd35dc1f: 16973.1triage_hosted_gpt_oss_120b_v1: 17021.1triage_hosted_gpt_oss_120b_v1 · rescored 5f2c057c: 17021.1triage_hosted_gpt_oss_120b_v1 · rescored 77adc97b: 17021.1triage_hosted_gpt_oss_120b_v1 · rescored cd35dc1f: 17021.1triage_local_x3_gpt_oss_20b_v1: 17235.0triage_local_x3_gpt_oss_20b_v1 · rescored 5f2c057c: 17235.0triage_local_x3_gpt_oss_20b_v1 · rescored 77adc97b: 17235.0triage_local_x3_gpt_oss_20b_v1 · rescored cd35dc1f: 17235.0triage_mistral_large_4_v1: 17290.2triage_mistral_large_4_v1 · rescored 5f2c057c: 17290.2triage_mistral_large_4_v1 · rescored 77adc97b: 17290.2triage_mistral_large_4_v1 · rescored cd35dc1f: 17290.2triage_hosted_qwen35_122b_v1: 17661.6triage_hosted_qwen35_122b_v1 · rescored 5f2c057c: 17661.6triage_hosted_qwen35_122b_v1 · rescored 77adc97b: 17661.6triage_hosted_qwen35_122b_v1 · rescored cd35dc1f: 17661.6triage_hosted_qwen3_30b_a3b_v1: 17976.1triage_hosted_qwen3_30b_a3b_v1 · rescored 3d3a6afd: 17976.1triage_hosted_qwen3_30b_a3b_v1 · rescored 5f2c057c: 17976.1triage_hosted_qwen3_30b_a3b_v1 · rescored 77adc97b: 17976.1triage_hosted_qwen3_30b_a3b_v1 · rescored cd35dc1f: 17976.1triage_local_mac_gpt_oss_20b_v1: 19952.6triage_local_mac_gpt_oss_20b_v1 · rescored 5f2c057c: 19952.6triage_local_mac_gpt_oss_20b_v1 · rescored 77adc97b: 19952.6triage_local_mac_gpt_oss_20b_v1 · rescored cd35dc1f: 19952.6triage_local_dgx_ministral_3_14b_v1: 20452.2triage_local_dgx_ministral_3_14b_v1 · rescored 3d3a6afd: 20452.2triage_local_dgx_ministral_3_14b_v1 · rescored 5f2c057c: 20452.2triage_local_dgx_ministral_3_14b_v1 · rescored 77adc97b: 20452.2triage_local_dgx_ministral_3_14b_v1 · rescored cd35dc1f: 20452.2triage_mistral_large_4_v2: 20489.0triage_mistral_large_4_v2 · rescored 3d3a6afd: 20489.0triage_mistral_large_4_v2 · rescored 5f2c057c: 20489.0triage_local_x3_qwen38_27b_v2: 20630.0triage_local_x3_qwen38_27b_v2 · rescored 3d3a6afd: 20630.0triage_local_x3_qwen38_27b_v2 · rescored 5f2c057c: 20630.0triage_local_x3_nemotron_35_bolt_e4b_v1 · repeats: 21266.2triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c, repeats: 21266.2triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 77adc97b, repeats: 21266.2triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f, repeats: 21266.2triage_kimi_k31_pro_v2: 21532.6triage_kimi_k31_pro_v2 · rescored 3d3a6afd: 21532.6triage_kimi_k31_pro_v2 · rescored 5f2c057c: 21532.6triage_local_mac_phi4_14b_v1: 21773.0triage_local_mac_phi4_14b_v1 · rescored 3d3a6afd: 21773.0triage_local_mac_phi4_14b_v1 · rescored 5f2c057c: 21773.0triage_local_mac_phi4_14b_v1 · rescored 77adc97b: 21773.0triage_local_mac_phi4_14b_v1 · rescored cd35dc1f: 21773.0triage_local_x3_ministral_3_14b_v1: 22057.2triage_local_x3_ministral_3_14b_v1 · rescored 3d3a6afd: 22057.2triage_local_x3_ministral_3_14b_v1 · rescored 5f2c057c: 22057.2triage_local_x3_ministral_3_14b_v1 · rescored 77adc97b: 22057.2triage_local_x3_ministral_3_14b_v1 · rescored cd35dc1f: 22057.2triage_local_x3_nemotron_35_bolt_e4b_v1: 22252.4triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c: 22252.4triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 77adc97b: 22252.4triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f: 22252.4triage_baseline_pro_v1: 22736.3triage_baseline_pro_v1 · rescored 3d3a6afd: 22736.3triage_baseline_pro_v1 · rescored 5f2c057c: 22736.3triage_baseline_pro_v1 · rescored 77adc97b: 22736.3triage_baseline_pro_v1 · rescored cd35dc1f: 22736.3triage_local_mac_nemotron_35_bolt_e4b_v1: 23126.4triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c: 23126.4triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored 77adc97b: 23126.4triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f: 23126.4triage_local_mac_gemma4_31b_v2: 25063.1triage_local_mac_gemma4_31b_v2 · rescored 3d3a6afd: 25063.1triage_local_mac_gemma4_31b_v2 · rescored 5f2c057c: 25063.1triage_baseline_pro_v1 · repeats ×3: 26284.7triage_baseline_pro_v1 · rescored 3d3a6afd, repeats ×3: 26284.7triage_baseline_pro_v1 · rescored 5f2c057c, repeats ×3: 26284.7triage_baseline_pro_v1 · rescored 77adc97b, repeats ×3: 26284.7triage_baseline_pro_v1 · rescored cd35dc1f, repeats ×3: 26284.7triage_local_mac_ministral_3_14b_v1: 27377.7triage_local_mac_ministral_3_14b_v1 · rescored 3d3a6afd: 27377.7triage_local_mac_ministral_3_14b_v1 · rescored 5f2c057c: 27377.7triage_local_mac_ministral_3_14b_v1 · rescored 77adc97b: 27377.7triage_local_mac_ministral_3_14b_v1 · rescored cd35dc1f: 27377.7triage_gemini_4_argon_v1: 30118.2triage_gemini_4_argon_v1 · rescored 3d3a6afd: 30118.2triage_gemini_4_argon_v1 · rescored 5f2c057c: 30118.2triage_gemini_4_argon_v1 · rescored 77adc97b: 30118.2triage_gemini_4_argon_v1 · rescored cd35dc1f: 30118.2triage_local_mac_qwen38_27b_v2: 33394.6triage_local_mac_qwen38_27b_v2 · rescored 3d3a6afd: 33394.6triage_local_mac_qwen38_27b_v2 · rescored 5f2c057c: 33394.6triage_local_dgx_qwen38_27b_v1: 37825.9triage_local_dgx_qwen38_27b_v1 · rescored 5f2c057c: 37825.9triage_local_dgx_qwen38_27b_v1 · rescored 77adc97b: 37825.9triage_local_dgx_qwen38_27b_v1 · rescored cd35dc1f: 37825.9triage_local_dgx_qwen35_122b_v1: 38007.3triage_local_dgx_qwen35_122b_v1 · rescored 5f2c057c: 38007.3triage_local_dgx_qwen35_122b_v1 · rescored 77adc97b: 38007.3triage_local_dgx_qwen35_122b_v1 · rescored cd35dc1f: 38007.3triage_local_x3_nemotron_35_lightning_30b_v1: 39458.3triage_local_x3_nemotron_35_lightning_30b_v1 · rescored 5f2c057c: 39458.3triage_local_x3_nemotron_35_lightning_30b_v1 · rescored 77adc97b: 39458.3triage_local_x3_nemotron_35_lightning_30b_v1 · rescored cd35dc1f: 39458.3triage_local_dgx_nemotron_35_lightning_30b_v1: 39979.1triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored 5f2c057c: 39979.1triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored 77adc97b: 39979.1triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored cd35dc1f: 39979.1triage_local_dgx_nemotron_35_super_120b_v1: 41018.6triage_local_dgx_nemotron_35_super_120b_v1 · rescored 5f2c057c: 41018.6triage_local_dgx_nemotron_35_super_120b_v1 · rescored 77adc97b: 41018.6triage_local_dgx_nemotron_35_super_120b_v1 · rescored cd35dc1f: 41018.6triage_kimi_k31_pro_v1: 46089.2triage_kimi_k31_pro_v1 · rescored 5f2c057c: 46089.2triage_kimi_k31_pro_v1 · rescored 77adc97b: 46089.2triage_kimi_k31_pro_v1 · rescored cd35dc1f: 46089.2triage_local_x3_qwen38_27b_v1: 46161.2triage_local_x3_qwen38_27b_v1 · rescored 5f2c057c: 46161.2triage_local_x3_qwen38_27b_v1 · rescored 77adc97b: 46161.2triage_local_x3_qwen38_27b_v1 · rescored cd35dc1f: 46161.2triage_local_dgx_gemma4_31b_v1: 58862.8triage_local_dgx_gemma4_31b_v1 · rescored 5f2c057c: 58862.8triage_local_dgx_gemma4_31b_v1 · rescored 77adc97b: 58862.8triage_local_dgx_gemma4_31b_v1 · rescored cd35dc1f: 58862.8triage_local_mac_nemotron_35_lightning_30b_v1: 66004.2triage_local_mac_nemotron_35_lightning_30b_v1 · rescored 5f2c057c: 66004.2triage_local_mac_nemotron_35_lightning_30b_v1 · rescored 77adc97b: 66004.2triage_local_mac_nemotron_35_lightning_30b_v1 · rescored cd35dc1f: 66004.2triage_local_mac_qwen38_27b_v1: 69674.0triage_local_mac_qwen38_27b_v1 · rescored 5f2c057c: 69674.0triage_local_mac_qwen38_27b_v1 · rescored 77adc97b: 69674.0triage_local_mac_qwen38_27b_v1 · rescored cd35dc1f: 69674.0triage_local_x3_gemma4_31b_v1: 72058.0triage_local_x3_gemma4_31b_v1 · rescored 5f2c057c: 72058.0triage_local_x3_gemma4_31b_v1 · rescored 77adc97b: 72058.0triage_local_x3_gemma4_31b_v1 · rescored cd35dc1f: 72058.0triage_local_mac_gemma4_31b_v1: 100888.9triage_local_mac_gemma4_31b_v1 · rescored 5f2c057c: 100888.9triage_local_mac_gemma4_31b_v1 · rescored 77adc97b: 100888.9triage_local_mac_gemma4_31b_v1 · rescored cd35dc1f: 100888.9triage_local_dgx_mistral_medium_35_128b_v1: 117538.1triage_local_dgx_mistral_medium_35_128b_v1 · rescored 3d3a6afd: 117538.1triage_local_dgx_mistral_medium_35_128b_v1 · rescored 5f2c057c: 117538.1triage_local_dgx_mistral_medium_35_128b_v1 · rescored 77adc97b: 117538.1triage_local_dgx_mistral_medium_35_128b_v1 · rescored cd35dc1f: 117538.10.09 · better117538.1

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean latency_msitemsscored against
triage_stub_escalate_v10.09065gold/orrery_triage_v1_labelled gold
triage_stub_escalate_v1 · rescored 3d3a6afd0.09065gold/orrery_triage_v1_labelled gold
triage_stub_escalate_v1 · rescored 5f2c057c0.09065gold/orrery_triage_v1_labelled gold
triage_stub_escalate_v1 · rescored 77adc97b0.09065gold/orrery_triage_v1_labelled gold
triage_stub_escalate_v1 · rescored cd35dc1f0.09065gold/orrery_triage_v1_labelled gold
triage_stub_dismiss_v10.10365gold/orrery_triage_v1_labelled gold
triage_stub_dismiss_v1 · rescored 3d3a6afd0.10365gold/orrery_triage_v1_labelled gold
triage_stub_dismiss_v1 · rescored 5f2c057c0.10365gold/orrery_triage_v1_labelled gold
triage_stub_dismiss_v1 · rescored 77adc97b0.10365gold/orrery_triage_v1_labelled gold
triage_stub_dismiss_v1 · rescored cd35dc1f0.10365gold/orrery_triage_v1_labelled gold
triage_deepseek_41flash_v23281.965gold/orrery_triage_v1_labelled gold
triage_deepseek_41flash_v2 · rescored 3d3a6afd3281.965gold/orrery_triage_v1_labelled gold
triage_deepseek_41flash_v2 · rescored 5f2c057c3281.965gold/orrery_triage_v1_labelled gold
triage_gemini_35_flash_lite_v13487.365gold/orrery_triage_v1_labelled gold
triage_gemini_35_flash_lite_v1 · rescored 3d3a6afd3487.365gold/orrery_triage_v1_labelled gold
triage_gemini_35_flash_lite_v1 · rescored 5f2c057c3487.365gold/orrery_triage_v1_labelled gold
triage_gemini_35_flash_lite_v1 · rescored 77adc97b3487.365gold/orrery_triage_v1_labelled gold
triage_gemini_35_flash_lite_v1 · rescored cd35dc1f3487.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_bolt_e4b_v23880.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afd3880.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_bolt_e4b_v2 · rescored 5f2c057c3880.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_20b_v24680.765gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_20b_v2 · rescored 3d3a6afd4680.765gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_20b_v2 · rescored 5f2c057c4680.765gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v25646.065gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afd5646.065gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v2 · rescored 5f2c057c5646.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_12b_noreason_v15697.465gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_12b_noreason_v1 · rescored 3d3a6afd5697.465gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_12b_noreason_v1 · rescored 5f2c057c5697.465gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_12b_noreason_v1 · rescored 77adc97b5697.465gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_12b_noreason_v1 · rescored cd35dc1f5697.465gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama31_8b_v15766.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama31_8b_v1 · rescored 3d3a6afd5766.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama31_8b_v1 · rescored 5f2c057c5766.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama31_8b_v1 · rescored 77adc97b5766.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama31_8b_v1 · rescored cd35dc1f5766.065gold/orrery_triage_v1_labelled gold
triage_local_mac_gpt_oss_20b_v25802.565gold/orrery_triage_v1_labelled gold
triage_local_mac_gpt_oss_20b_v2 · rescored 3d3a6afd5802.565gold/orrery_triage_v1_labelled gold
triage_local_mac_gpt_oss_20b_v2 · rescored 5f2c057c5802.565gold/orrery_triage_v1_labelled gold
triage_local_x3_llama31_8b_v15836.865gold/orrery_triage_v1_labelled gold
triage_local_x3_llama31_8b_v1 · rescored 3d3a6afd5836.865gold/orrery_triage_v1_labelled gold
triage_local_x3_llama31_8b_v1 · rescored 5f2c057c5836.865gold/orrery_triage_v1_labelled gold
triage_local_x3_llama31_8b_v1 · rescored 77adc97b5836.865gold/orrery_triage_v1_labelled gold
triage_local_x3_llama31_8b_v1 · rescored cd35dc1f5836.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen3_30b_a3b_v15859.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen3_30b_a3b_v1 · rescored 3d3a6afd5859.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen3_30b_a3b_v1 · rescored 5f2c057c5859.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen3_30b_a3b_v1 · rescored 77adc97b5859.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen3_30b_a3b_v1 · rescored cd35dc1f5859.165gold/orrery_triage_v1_labelled gold
triage_local_x3_gpt_oss_20b_v25863.565gold/orrery_triage_v1_labelled gold
triage_local_x3_gpt_oss_20b_v2 · rescored 3d3a6afd5863.565gold/orrery_triage_v1_labelled gold
triage_local_x3_gpt_oss_20b_v2 · rescored 5f2c057c5863.565gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_bolt_e4b_v25883.165gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_bolt_e4b_v2 · rescored 3d3a6afd5883.165gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_bolt_e4b_v2 · rescored 5f2c057c5883.165gold/orrery_triage_v1_labelled gold
triage_hosted_mistral_medium_35_v16103.365gold/orrery_triage_v1_labelled gold
triage_hosted_mistral_medium_35_v1 · rescored 3d3a6afd6103.365gold/orrery_triage_v1_labelled gold
triage_hosted_mistral_medium_35_v1 · rescored 5f2c057c6103.365gold/orrery_triage_v1_labelled gold
triage_hosted_mistral_medium_35_v1 · rescored 77adc97b6103.365gold/orrery_triage_v1_labelled gold
triage_hosted_mistral_medium_35_v1 · rescored cd35dc1f6103.365gold/orrery_triage_v1_labelled gold
triage_hosted_gpt_oss_120b_v26540.765gold/orrery_triage_v1_labelled gold
triage_hosted_gpt_oss_120b_v2 · rescored 3d3a6afd6540.765gold/orrery_triage_v1_labelled gold
triage_hosted_gpt_oss_120b_v2 · rescored 5f2c057c6540.765gold/orrery_triage_v1_labelled gold
triage_hosted_qwen35_122b_v26681.465gold/orrery_triage_v1_labelled gold
triage_hosted_qwen35_122b_v2 · rescored 3d3a6afd6681.465gold/orrery_triage_v1_labelled gold
triage_hosted_qwen35_122b_v2 · rescored 5f2c057c6681.465gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_lightning_30b_v26785.765gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_lightning_30b_v2 · rescored 3d3a6afd6785.765gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_lightning_30b_v2 · rescored 5f2c057c6785.765gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen3_30b_a3b_v16932.365gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen3_30b_a3b_v1 · rescored 3d3a6afd6932.365gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen3_30b_a3b_v1 · rescored 5f2c057c6932.365gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen3_30b_a3b_v1 · rescored 77adc97b6932.365gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen3_30b_a3b_v1 · rescored cd35dc1f6932.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_120b_v27391.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_120b_v2 · rescored 3d3a6afd7391.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_120b_v2 · rescored 5f2c057c7391.165gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · repeats ×37528.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored 3d3a6afd, repeats ×37528.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored 5f2c057c, repeats ×37528.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored 77adc97b, repeats ×37528.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored cd35dc1f, repeats ×37528.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v17625.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored 3d3a6afd7625.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored 5f2c057c7625.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored 77adc97b7625.965gold/orrery_triage_v1_labelled gold
triage_haiku_55_v1 · rescored cd35dc1f7625.965gold/orrery_triage_v1_labelled gold
triage_local_mac_llama31_8b_v17780.065gold/orrery_triage_v1_labelled gold
triage_local_mac_llama31_8b_v1 · rescored 3d3a6afd7780.065gold/orrery_triage_v1_labelled gold
triage_local_mac_llama31_8b_v1 · rescored 5f2c057c7780.065gold/orrery_triage_v1_labelled gold
triage_local_mac_llama31_8b_v1 · rescored 77adc97b7780.065gold/orrery_triage_v1_labelled gold
triage_local_mac_llama31_8b_v1 · rescored cd35dc1f7780.065gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_lightning_30b_v27797.365gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_lightning_30b_v2 · rescored 3d3a6afd7797.365gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_lightning_30b_v2 · rescored 5f2c057c7797.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v17824.565gold/orrery_triage_v1_labelled gold
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 3d3a6afd7824.565gold/orrery_triage_v1_labelled gold
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 5f2c057c7824.565gold/orrery_triage_v1_labelled gold
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored 77adc97b7824.565gold/orrery_triage_v1_labelled gold
triage_local_dgx_vllm_qwen3_30b_a3b_nvfp4_v1 · rescored cd35dc1f7824.565gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v27973.365gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v2 · rescored 3d3a6afd7973.365gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v2 · rescored 5f2c057c7973.365gold/orrery_triage_v1_labelled gold
triage_baseline_pro_noreason_v18352.265gold/orrery_triage_v1_labelled gold
triage_baseline_pro_noreason_v1 · rescored 3d3a6afd8352.265gold/orrery_triage_v1_labelled gold
triage_baseline_pro_noreason_v1 · rescored 5f2c057c8352.265gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen3_30b_a3b_v18996.065gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen3_30b_a3b_v1 · rescored 3d3a6afd8996.065gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen3_30b_a3b_v1 · rescored 5f2c057c8996.065gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen3_30b_a3b_v1 · rescored 77adc97b8996.065gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen3_30b_a3b_v1 · rescored cd35dc1f8996.065gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_12b_noreason_v19174.265gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_12b_noreason_v1 · rescored 3d3a6afd9174.265gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_12b_noreason_v1 · rescored 5f2c057c9174.265gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_12b_noreason_v1 · rescored 77adc97b9174.265gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_12b_noreason_v1 · rescored cd35dc1f9174.265gold/orrery_triage_v1_labelled gold
triage_deepseek_41flash_v19832.865gold/orrery_triage_v1_labelled gold
triage_deepseek_41flash_v1 · rescored 5f2c057c9832.865gold/orrery_triage_v1_labelled gold
triage_deepseek_41flash_v1 · rescored 77adc97b9832.865gold/orrery_triage_v1_labelled gold
triage_deepseek_41flash_v1 · rescored cd35dc1f9832.865gold/orrery_triage_v1_labelled gold
triage_hosted_gemma4_31b_v110510.365gold/orrery_triage_v1_labelled gold
triage_hosted_gemma4_31b_v1 · rescored 5f2c057c10510.365gold/orrery_triage_v1_labelled gold
triage_hosted_gemma4_31b_v1 · rescored 77adc97b10510.365gold/orrery_triage_v1_labelled gold
triage_hosted_gemma4_31b_v1 · rescored cd35dc1f10510.365gold/orrery_triage_v1_labelled gold
triage_hosted_gemma4_31b_v210549.765gold/orrery_triage_v1_labelled gold
triage_hosted_gemma4_31b_v2 · rescored 3d3a6afd10549.765gold/orrery_triage_v1_labelled gold
triage_hosted_gemma4_31b_v2 · rescored 5f2c057c10549.765gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_lightning_30b_v211064.965gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_lightning_30b_v2 · rescored 3d3a6afd11064.965gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_lightning_30b_v2 · rescored 5f2c057c11064.965gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v1 · repeats ×311597.765gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v1 · rescored 5f2c057c, repeats ×311597.765gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v1 · rescored 77adc97b, repeats ×311597.765gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v1 · rescored cd35dc1f, repeats ×311597.765gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · repeats ×312005.565gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored 3d3a6afd, repeats ×312005.565gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored 5f2c057c, repeats ×312005.565gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored 77adc97b, repeats ×312005.565gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored cd35dc1f, repeats ×312005.565gold/orrery_triage_v1_labelled gold
triage_sonnet_55_v112035.065gold/orrery_triage_v1_labelled gold
triage_sonnet_55_v1 · rescored 3d3a6afd12035.065gold/orrery_triage_v1_labelled gold
triage_sonnet_55_v1 · rescored 5f2c057c12035.065gold/orrery_triage_v1_labelled gold
triage_sonnet_55_v1 · rescored 77adc97b12035.065gold/orrery_triage_v1_labelled gold
triage_sonnet_55_v1 · rescored cd35dc1f12035.065gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v112200.565gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v1 · rescored 5f2c057c12200.565gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v1 · rescored 77adc97b12200.565gold/orrery_triage_v1_labelled gold
triage_gpt_6_luna_v1 · rescored cd35dc1f12200.565gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v112241.465gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored 3d3a6afd12241.465gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored 5f2c057c12241.465gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored 77adc97b12241.465gold/orrery_triage_v1_labelled gold
triage_gpt_61_sol_v1 · rescored cd35dc1f12241.465gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_31b_v213030.565gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_31b_v2 · rescored 3d3a6afd13030.565gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_31b_v2 · rescored 5f2c057c13030.565gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama4_scout_v113354.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama4_scout_v1 · rescored 3d3a6afd13354.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama4_scout_v1 · rescored 5f2c057c13354.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama4_scout_v1 · rescored 77adc97b13354.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_llama4_scout_v1 · rescored cd35dc1f13354.065gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_20b_v113455.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_20b_v1 · rescored 5f2c057c13455.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_20b_v1 · rescored 77adc97b13455.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_20b_v1 · rescored cd35dc1f13455.965gold/orrery_triage_v1_labelled gold
triage_local_x3_gemma4_31b_v214320.165gold/orrery_triage_v1_labelled gold
triage_local_x3_gemma4_31b_v2 · rescored 3d3a6afd14320.165gold/orrery_triage_v1_labelled gold
triage_local_x3_gemma4_31b_v2 · rescored 5f2c057c14320.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen35_122b_v214511.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen35_122b_v2 · rescored 3d3a6afd14511.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen35_122b_v2 · rescored 5f2c057c14511.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_120b_v114765.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_120b_v1 · rescored 5f2c057c14765.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_120b_v1 · rescored 77adc97b14765.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_gpt_oss_120b_v1 · rescored cd35dc1f14765.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_bolt_e4b_v115258.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c15258.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored 77adc97b15258.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f15258.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_super_120b_v215897.765gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_super_120b_v2 · rescored 3d3a6afd15897.765gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_super_120b_v2 · rescored 5f2c057c15897.765gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen38_27b_v216558.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen38_27b_v2 · rescored 3d3a6afd16558.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen38_27b_v2 · rescored 5f2c057c16558.965gold/orrery_triage_v1_labelled gold
triage_local_x3_phi4_14b_v116911.265gold/orrery_triage_v1_labelled gold
triage_local_x3_phi4_14b_v1 · rescored 3d3a6afd16911.265gold/orrery_triage_v1_labelled gold
triage_local_x3_phi4_14b_v1 · rescored 5f2c057c16911.265gold/orrery_triage_v1_labelled gold
triage_local_x3_phi4_14b_v1 · rescored 77adc97b16911.265gold/orrery_triage_v1_labelled gold
triage_local_x3_phi4_14b_v1 · rescored cd35dc1f16911.265gold/orrery_triage_v1_labelled gold
triage_local_dgx_phi4_14b_v116973.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_phi4_14b_v1 · rescored 3d3a6afd16973.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_phi4_14b_v1 · rescored 5f2c057c16973.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_phi4_14b_v1 · rescored 77adc97b16973.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_phi4_14b_v1 · rescored cd35dc1f16973.165gold/orrery_triage_v1_labelled gold
triage_hosted_gpt_oss_120b_v117021.165gold/orrery_triage_v1_labelled gold
triage_hosted_gpt_oss_120b_v1 · rescored 5f2c057c17021.165gold/orrery_triage_v1_labelled gold
triage_hosted_gpt_oss_120b_v1 · rescored 77adc97b17021.165gold/orrery_triage_v1_labelled gold
triage_hosted_gpt_oss_120b_v1 · rescored cd35dc1f17021.165gold/orrery_triage_v1_labelled gold
triage_local_x3_gpt_oss_20b_v117235.065gold/orrery_triage_v1_labelled gold
triage_local_x3_gpt_oss_20b_v1 · rescored 5f2c057c17235.065gold/orrery_triage_v1_labelled gold
triage_local_x3_gpt_oss_20b_v1 · rescored 77adc97b17235.065gold/orrery_triage_v1_labelled gold
triage_local_x3_gpt_oss_20b_v1 · rescored cd35dc1f17235.065gold/orrery_triage_v1_labelled gold
triage_mistral_large_4_v117290.265gold/orrery_triage_v1_labelled gold
triage_mistral_large_4_v1 · rescored 5f2c057c17290.265gold/orrery_triage_v1_labelled gold
triage_mistral_large_4_v1 · rescored 77adc97b17290.265gold/orrery_triage_v1_labelled gold
triage_mistral_large_4_v1 · rescored cd35dc1f17290.265gold/orrery_triage_v1_labelled gold
triage_hosted_qwen35_122b_v117661.665gold/orrery_triage_v1_labelled gold
triage_hosted_qwen35_122b_v1 · rescored 5f2c057c17661.665gold/orrery_triage_v1_labelled gold
triage_hosted_qwen35_122b_v1 · rescored 77adc97b17661.665gold/orrery_triage_v1_labelled gold
triage_hosted_qwen35_122b_v1 · rescored cd35dc1f17661.665gold/orrery_triage_v1_labelled gold
triage_hosted_qwen3_30b_a3b_v117976.165gold/orrery_triage_v1_labelled gold
triage_hosted_qwen3_30b_a3b_v1 · rescored 3d3a6afd17976.165gold/orrery_triage_v1_labelled gold
triage_hosted_qwen3_30b_a3b_v1 · rescored 5f2c057c17976.165gold/orrery_triage_v1_labelled gold
triage_hosted_qwen3_30b_a3b_v1 · rescored 77adc97b17976.165gold/orrery_triage_v1_labelled gold
triage_hosted_qwen3_30b_a3b_v1 · rescored cd35dc1f17976.165gold/orrery_triage_v1_labelled gold
triage_local_mac_gpt_oss_20b_v119952.665gold/orrery_triage_v1_labelled gold
triage_local_mac_gpt_oss_20b_v1 · rescored 5f2c057c19952.665gold/orrery_triage_v1_labelled gold
triage_local_mac_gpt_oss_20b_v1 · rescored 77adc97b19952.665gold/orrery_triage_v1_labelled gold
triage_local_mac_gpt_oss_20b_v1 · rescored cd35dc1f19952.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_ministral_3_14b_v120452.265gold/orrery_triage_v1_labelled gold
triage_local_dgx_ministral_3_14b_v1 · rescored 3d3a6afd20452.265gold/orrery_triage_v1_labelled gold
triage_local_dgx_ministral_3_14b_v1 · rescored 5f2c057c20452.265gold/orrery_triage_v1_labelled gold
triage_local_dgx_ministral_3_14b_v1 · rescored 77adc97b20452.265gold/orrery_triage_v1_labelled gold
triage_local_dgx_ministral_3_14b_v1 · rescored cd35dc1f20452.265gold/orrery_triage_v1_labelled gold
triage_mistral_large_4_v220489.065gold/orrery_triage_v1_labelled gold
triage_mistral_large_4_v2 · rescored 3d3a6afd20489.065gold/orrery_triage_v1_labelled gold
triage_mistral_large_4_v2 · rescored 5f2c057c20489.065gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen38_27b_v220630.065gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen38_27b_v2 · rescored 3d3a6afd20630.065gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen38_27b_v2 · rescored 5f2c057c20630.065gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v1 · repeats21266.265gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c, repeats21266.265gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 77adc97b, repeats21266.265gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f, repeats21266.265gold/orrery_triage_v1_labelled gold
triage_kimi_k31_pro_v221532.665gold/orrery_triage_v1_labelled gold
triage_kimi_k31_pro_v2 · rescored 3d3a6afd21532.665gold/orrery_triage_v1_labelled gold
triage_kimi_k31_pro_v2 · rescored 5f2c057c21532.665gold/orrery_triage_v1_labelled gold
triage_local_mac_phi4_14b_v121773.065gold/orrery_triage_v1_labelled gold
triage_local_mac_phi4_14b_v1 · rescored 3d3a6afd21773.065gold/orrery_triage_v1_labelled gold
triage_local_mac_phi4_14b_v1 · rescored 5f2c057c21773.065gold/orrery_triage_v1_labelled gold
triage_local_mac_phi4_14b_v1 · rescored 77adc97b21773.065gold/orrery_triage_v1_labelled gold
triage_local_mac_phi4_14b_v1 · rescored cd35dc1f21773.065gold/orrery_triage_v1_labelled gold
triage_local_x3_ministral_3_14b_v122057.265gold/orrery_triage_v1_labelled gold
triage_local_x3_ministral_3_14b_v1 · rescored 3d3a6afd22057.265gold/orrery_triage_v1_labelled gold
triage_local_x3_ministral_3_14b_v1 · rescored 5f2c057c22057.265gold/orrery_triage_v1_labelled gold
triage_local_x3_ministral_3_14b_v1 · rescored 77adc97b22057.265gold/orrery_triage_v1_labelled gold
triage_local_x3_ministral_3_14b_v1 · rescored cd35dc1f22057.265gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v122252.465gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c22252.465gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored 77adc97b22252.465gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f22252.465gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v122736.365gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored 3d3a6afd22736.365gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored 5f2c057c22736.365gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored 77adc97b22736.365gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored cd35dc1f22736.365gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_bolt_e4b_v123126.465gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored 5f2c057c23126.465gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored 77adc97b23126.465gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_bolt_e4b_v1 · rescored cd35dc1f23126.465gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_31b_v225063.165gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_31b_v2 · rescored 3d3a6afd25063.165gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_31b_v2 · rescored 5f2c057c25063.165gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · repeats ×326284.765gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored 3d3a6afd, repeats ×326284.765gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored 5f2c057c, repeats ×326284.765gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored 77adc97b, repeats ×326284.765gold/orrery_triage_v1_labelled gold
triage_baseline_pro_v1 · rescored cd35dc1f, repeats ×326284.765gold/orrery_triage_v1_labelled gold
triage_local_mac_ministral_3_14b_v127377.765gold/orrery_triage_v1_labelled gold
triage_local_mac_ministral_3_14b_v1 · rescored 3d3a6afd27377.765gold/orrery_triage_v1_labelled gold
triage_local_mac_ministral_3_14b_v1 · rescored 5f2c057c27377.765gold/orrery_triage_v1_labelled gold
triage_local_mac_ministral_3_14b_v1 · rescored 77adc97b27377.765gold/orrery_triage_v1_labelled gold
triage_local_mac_ministral_3_14b_v1 · rescored cd35dc1f27377.765gold/orrery_triage_v1_labelled gold
triage_gemini_4_argon_v130118.265gold/orrery_triage_v1_labelled gold
triage_gemini_4_argon_v1 · rescored 3d3a6afd30118.265gold/orrery_triage_v1_labelled gold
triage_gemini_4_argon_v1 · rescored 5f2c057c30118.265gold/orrery_triage_v1_labelled gold
triage_gemini_4_argon_v1 · rescored 77adc97b30118.265gold/orrery_triage_v1_labelled gold
triage_gemini_4_argon_v1 · rescored cd35dc1f30118.265gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen38_27b_v233394.665gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen38_27b_v2 · rescored 3d3a6afd33394.665gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen38_27b_v2 · rescored 5f2c057c33394.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen38_27b_v137825.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen38_27b_v1 · rescored 5f2c057c37825.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen38_27b_v1 · rescored 77adc97b37825.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen38_27b_v1 · rescored cd35dc1f37825.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen35_122b_v138007.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen35_122b_v1 · rescored 5f2c057c38007.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen35_122b_v1 · rescored 77adc97b38007.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_qwen35_122b_v1 · rescored cd35dc1f38007.365gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_lightning_30b_v139458.365gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_lightning_30b_v1 · rescored 5f2c057c39458.365gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_lightning_30b_v1 · rescored 77adc97b39458.365gold/orrery_triage_v1_labelled gold
triage_local_x3_nemotron_35_lightning_30b_v1 · rescored cd35dc1f39458.365gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_lightning_30b_v139979.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored 5f2c057c39979.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored 77adc97b39979.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_lightning_30b_v1 · rescored cd35dc1f39979.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_super_120b_v141018.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_super_120b_v1 · rescored 5f2c057c41018.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_super_120b_v1 · rescored 77adc97b41018.665gold/orrery_triage_v1_labelled gold
triage_local_dgx_nemotron_35_super_120b_v1 · rescored cd35dc1f41018.665gold/orrery_triage_v1_labelled gold
triage_kimi_k31_pro_v146089.265gold/orrery_triage_v1_labelled gold
triage_kimi_k31_pro_v1 · rescored 5f2c057c46089.265gold/orrery_triage_v1_labelled gold
triage_kimi_k31_pro_v1 · rescored 77adc97b46089.265gold/orrery_triage_v1_labelled gold
triage_kimi_k31_pro_v1 · rescored cd35dc1f46089.265gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen38_27b_v146161.265gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen38_27b_v1 · rescored 5f2c057c46161.265gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen38_27b_v1 · rescored 77adc97b46161.265gold/orrery_triage_v1_labelled gold
triage_local_x3_qwen38_27b_v1 · rescored cd35dc1f46161.265gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_31b_v158862.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_31b_v1 · rescored 5f2c057c58862.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_31b_v1 · rescored 77adc97b58862.865gold/orrery_triage_v1_labelled gold
triage_local_dgx_gemma4_31b_v1 · rescored cd35dc1f58862.865gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_lightning_30b_v166004.265gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_lightning_30b_v1 · rescored 5f2c057c66004.265gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_lightning_30b_v1 · rescored 77adc97b66004.265gold/orrery_triage_v1_labelled gold
triage_local_mac_nemotron_35_lightning_30b_v1 · rescored cd35dc1f66004.265gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen38_27b_v169674.065gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen38_27b_v1 · rescored 5f2c057c69674.065gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen38_27b_v1 · rescored 77adc97b69674.065gold/orrery_triage_v1_labelled gold
triage_local_mac_qwen38_27b_v1 · rescored cd35dc1f69674.065gold/orrery_triage_v1_labelled gold
triage_local_x3_gemma4_31b_v172058.065gold/orrery_triage_v1_labelled gold
triage_local_x3_gemma4_31b_v1 · rescored 5f2c057c72058.065gold/orrery_triage_v1_labelled gold
triage_local_x3_gemma4_31b_v1 · rescored 77adc97b72058.065gold/orrery_triage_v1_labelled gold
triage_local_x3_gemma4_31b_v1 · rescored cd35dc1f72058.065gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_31b_v1100888.965gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_31b_v1 · rescored 5f2c057c100888.965gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_31b_v1 · rescored 77adc97b100888.965gold/orrery_triage_v1_labelled gold
triage_local_mac_gemma4_31b_v1 · rescored cd35dc1f100888.965gold/orrery_triage_v1_labelled gold
triage_local_dgx_mistral_medium_35_128b_v1117538.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_mistral_medium_35_128b_v1 · rescored 3d3a6afd117538.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_mistral_medium_35_128b_v1 · rescored 5f2c057c117538.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_mistral_medium_35_128b_v1 · rescored 77adc97b117538.165gold/orrery_triage_v1_labelled gold
triage_local_dgx_mistral_medium_35_128b_v1 · rescored cd35dc1f117538.165gold/orrery_triage_v1_labelled gold

10on scifact_200

gold/scifact_200sf_first_k_n200_v1: 0.253sf_random_n200_v1: 1.631sf_minilm_n200_v1: 17.2sf_bm25_n200_v1: 22.3sf_bge_small_n200_v1: 29.4sf_e5_base_n200_v1: 59.4sf_mpnet_n200_v1: 65.4sf_deepseek_m_n200_v1: 1145.0sf_glm_s_bm25first_n200_v1 · pair: 3023.9sf_glm_s_n200_v1: 3119.5sf_glm_s_e5first_n200_v1 · pair: 3126.9sf_deepseek_s_n200_v1: 3452.8sf_llama_m_n200_v1: 4064.0sf_llama_l_n200_v1: 4553.7sf_llama_s_n200_v1 · repeats: 4873.8sf_qwen_s_n200_v1: 5650.6sf_gemma_m_n200_v1: 5663.7sf_qwen_s_n200_v1 · repeats: 5976.0sf_qwen_s_e5first_n200_v1 · pair: 6145.4sf_qwen_s_bm25first_n200_v1 · pair: 6196.2sf_glm_m_n200_v1: 6920.1sf_gemma_s_n200_v1: 7071.1sf_mistral_s_n200_v1: 7223.8sf_gemma_l_n200_v1: 9701.0sf_llama_l_n200_v1 · repeats: 11385.9sf_llama_s_n200_v1: 17481.60.253 · better17481.6

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

11on smoke_v1

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean latency_msitemsscored against
demo_v10.0095silver/smoke_v1 silver

12on summeval_100

gold/summeval_100se_random_n100_pick_v1: 0.044se_random_n100_pairwise_v1: 0.073se_random_n100_v1: 0.131se_length_n100_v1: 0.192se_rouge2_n100_v1: 180.9se_rouge1_n100_v1: 195.6se_rougeL_n100_v1: 274.3se_coverage_n100_v1: 892.0se_deepseek_m_n100_v1: 1120.3se_qwen_s_n100_pick_v1: 1492.1se_openai_s_n100_v1: 1550.9se_gemma_l_n100_pick_v1: 1813.0se_mistral_m_n100_v1: 1945.6se_openai_l_n100_v1: 2227.9se_anthropic_s_n100_v1: 2432.6se_gemma_l_pubtext_n100_v1 · pair: 2614.2se_openai_m_n100_v1: 2689.5se_anthropic_m_n100_v1: 3008.2se_gemma_l_pubtextcw_n100_v1 · pair: 3268.1se_anthropic_l_n100_v1: 3426.6se_glm_l_n100_v1: 3589.8se_glm_m_n100_v1: 3608.7se_qwen_s_n100_v1: 3953.6se_deepseek_s_n100_v1: 3984.8se_gemma_m_n100_v1: 4308.1se_qwen37max_n100_v1: 4514.3se_llama_m_n100_v1: 4834.2se_deepseek_l_n100_v1: 5535.2se_glm_s_n100_v1: 5617.5se_gemma_l_n100_v1: 6670.3se_qwen_m_n100_v1: 6770.4se_gemma_l_n100_pairwise_v1: 6980.8se_llama_s_n100_v1: 7141.3se_mistral_s_n100_v1: 9184.9se_mistral_l_n100_v1: 10884.9se_gemma_s_n100_v1: 11572.0se_llama_l_n100_v1: 12345.9se_bartscore_n100_v1: 15662.3se_qwen_s_n100_pairwise_v1: 20081.2se_unieval_n100_v1: 41676.1se_unievalpub_n100_v1: 44694.40.044 · better44694.4

scored against silverscored by a judgewrote the reference it is scored againstscored against a label set, not a reference setunmarked rows are scored against gold

armmean latency_msitemsscored against
se_random_n100_pick_v10.044100gold/summeval_100 gold
se_random_n100_pairwise_v10.073100gold/summeval_100 gold
se_random_n100_v10.131100gold/summeval_100 gold
se_length_n100_v10.192100gold/summeval_100 gold
se_rouge2_n100_v1180.9100gold/summeval_100 gold
se_rouge1_n100_v1195.6100gold/summeval_100 gold
se_rougeL_n100_v1274.3100gold/summeval_100 gold
se_coverage_n100_v1892.0100gold/summeval_100 gold
se_deepseek_m_n100_v11120.3100gold/summeval_100 gold
se_qwen_s_n100_pick_v11492.1100gold/summeval_100 gold
se_openai_s_n100_v11550.9100gold/summeval_100 gold
se_gemma_l_n100_pick_v11813.0100gold/summeval_100 gold
se_mistral_m_n100_v11945.6100gold/summeval_100 gold
se_openai_l_n100_v12227.9100gold/summeval_100 gold
se_anthropic_s_n100_v12432.6100gold/summeval_100 gold
se_gemma_l_pubtext_n100_v1 · pair2614.2100gold/summeval_100 gold
se_openai_m_n100_v12689.5100gold/summeval_100 gold
se_anthropic_m_n100_v13008.2100gold/summeval_100 gold
se_gemma_l_pubtextcw_n100_v1 · pair3268.1100gold/summeval_100 gold
se_anthropic_l_n100_v13426.6100gold/summeval_100 gold
se_glm_l_n100_v13589.8100gold/summeval_100 gold
se_glm_m_n100_v13608.7100gold/summeval_100 gold
se_qwen_s_n100_v13953.6100gold/summeval_100 gold
se_deepseek_s_n100_v13984.8100gold/summeval_100 gold
se_gemma_m_n100_v14308.1100gold/summeval_100 gold
se_qwen37max_n100_v14514.3100gold/summeval_100 gold
se_llama_m_n100_v14834.2100gold/summeval_100 gold
se_deepseek_l_n100_v15535.2100gold/summeval_100 gold
se_glm_s_n100_v15617.5100gold/summeval_100 gold
se_gemma_l_n100_v16670.3100gold/summeval_100 gold
se_qwen_m_n100_v16770.4100gold/summeval_100 gold
se_gemma_l_n100_pairwise_v16980.8100gold/summeval_100 gold
se_llama_s_n100_v17141.3100gold/summeval_100 gold
se_mistral_s_n100_v19184.9100gold/summeval_100 gold
se_mistral_l_n100_v110884.9100gold/summeval_100 gold
se_gemma_s_n100_v111572.0100gold/summeval_100 gold
se_llama_l_n100_v112345.9100gold/summeval_100 gold
se_bartscore_n100_v115662.3100gold/summeval_100 gold
se_qwen_s_n100_pairwise_v120081.2100gold/summeval_100 gold
se_unieval_n100_v141676.1100gold/summeval_100 gold
se_unievalpub_n100_v144694.4100gold/summeval_100 gold