Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.
Arena text preference / chinesehuman preference · 2026-10-15 · rating
fastchat-t5-3b
718.86 rating
Confidence interval: 681.57 – 756.16
224 votes
Category
chinese
Evaluated configuration
fastchat-t5-3b
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-15 · rating
fastchat-t5-3b
763.71 rating
Confidence interval: 734.31 – 793.11
428 votes
Category
coding
Evaluated configuration
fastchat-t5-3b
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating
fastchat-t5-3b
901.98 rating
Confidence interval: 878.65 – 925.31
744 votes
Category
creative_writing
Evaluated configuration
fastchat-t5-3b
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating
fastchat-t5-3b
934.53 rating
Confidence interval: 921.25 – 947.81
3442 votes
Category
english
Evaluated configuration
fastchat-t5-3b
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating
fastchat-t5-3b
634.89 rating
Confidence interval: 616.19 – 653.58
2836 votes
Category
exclude_ties
Evaluated configuration
fastchat-t5-3b
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating
fastchat-t5-3b
814.08 rating
Confidence interval: 793.39 – 834.77
797 votes
Category
hard_prompts
Evaluated configuration
fastchat-t5-3b
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating
fastchat-t5-3b
829.62 rating
Confidence interval: 807.03 – 852.20
690 votes
Category
hard_prompts_english
Evaluated configuration
fastchat-t5-3b
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating