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 / codinghuman preference · 2026-10-15 · rating
guanaco-33b
995.34 rating
Confidence interval: 960.81 – 1,029.88
263 votes
Category
coding
Evaluated configuration
guanaco-33b
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-15 · rating
guanaco-33b
1,079.84 rating
Confidence interval: 1,052.64 – 1,107.05
515 votes
Category
creative_writing
Evaluated configuration
guanaco-33b
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-15 · rating
guanaco-33b
1,085.31 rating
Confidence interval: 1,071.90 – 1,098.73
2352 votes
Category
english
Evaluated configuration
guanaco-33b
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-15 · rating
guanaco-33b
871.94 rating
Confidence interval: 854.31 – 889.57
2032 votes
Category
exclude_ties
Evaluated configuration
guanaco-33b
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-15 · rating
guanaco-33b
1,004.90 rating
Confidence interval: 980.36 – 1,029.45
517 votes
Category
hard_prompts
Evaluated configuration
guanaco-33b
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-15 · rating
guanaco-33b
1,020.13 rating
Confidence interval: 992.62 – 1,047.64
434 votes
Category
hard_prompts_english
Evaluated configuration
guanaco-33b
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-15 · rating