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 · 2026-10-28 · rating
Gemini Pro
1,096.35 rating
Confidence interval: 1,059.31 – 1,133.38
236 votes
Category
chinese
Evaluated configuration
Gemini Pro
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating
Gemini Pro
1,107.61 rating
Confidence interval: 1,084.53 – 1,130.68
678 votes
Category
coding
Evaluated configuration
Gemini Pro
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating
Gemini Pro
1,103.56 rating
Confidence interval: 1,082.84 – 1,124.27
953 votes
Category
creative_writing
Evaluated configuration
Gemini Pro
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating
Gemini Pro
1,153.71 rating
Confidence interval: 1,141.13 – 1,166.28
5086 votes
Category
english
Evaluated configuration
Gemini Pro
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating
Gemini Pro
977.49 rating
Confidence interval: 961.56 – 993.42
4502 votes
Category
exclude_ties
Evaluated configuration
Gemini Pro
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating
Gemini Pro
1,093.69 rating
Confidence interval: 1,076.30 – 1,111.08
1407 votes
Category
hard_prompts
Evaluated configuration
Gemini Pro
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating
Gemini Pro
1,110.62 rating
Confidence interval: 1,091.87 – 1,129.36
1189 votes
Category
hard_prompts_english
Evaluated configuration
Gemini Pro
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating