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
Koala 13B
880.86 rating
Confidence interval: 849.96 – 911.75
374 votes
Category
chinese
Evaluated configuration
Koala 13B
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating
Koala 13B
944.62 rating
Confidence interval: 921.18 – 968.06
747 votes
Category
coding
Evaluated configuration
Koala 13B
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating
Koala 13B
972.67 rating
Confidence interval: 952.43 – 992.91
1200 votes
Category
creative_writing
Evaluated configuration
Koala 13B
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating
Koala 13B
1,021.78 rating
Confidence interval: 1,010.78 – 1,032.78
5711 votes
Category
english
Evaluated configuration
Koala 13B
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating
Koala 13B
781.97 rating
Confidence interval: 767.18 – 796.76
4743 votes
Category
exclude_ties
Evaluated configuration
Koala 13B
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating
Koala 13B
930.30 rating
Confidence interval: 912.63 – 947.98
1250 votes
Category
hard_prompts
Evaluated configuration
Koala 13B
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating
Koala 13B
943.07 rating
Confidence interval: 923.86 – 962.28
1100 votes
Category
hard_prompts_english
Evaluated configuration
Koala 13B
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating