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
Reka Flash 21B (20240226)
1,135.27 rating
Confidence interval: 1,121.77 – 1,148.78
2774 votes
Category
chinese
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / codinghuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,153.37 rating
Confidence interval: 1,142.77 – 1,163.98
4748 votes
Category
coding
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / creative_writinghuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,123.54 rating
Confidence interval: 1,111.42 – 1,135.67
3490 votes
Category
creative_writing
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / englishhuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,179.06 rating
Confidence interval: 1,171.72 – 1,186.40
14444 votes
Category
english
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / exclude_tieshuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,031.91 rating
Confidence interval: 1,022.72 – 1,041.10
16076 votes
Category
exclude_ties
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / experthuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,106.98 rating
Confidence interval: 1,089.87 – 1,124.10
1325 votes
Category
expert
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / frenchhuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,174.70 rating
Confidence interval: 1,143.27 – 1,206.13
430 votes
Category
french
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / germanhuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,152.66 rating
Confidence interval: 1,130.87 – 1,174.45
798 votes
Category
german
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / hard_promptshuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,136.40 rating
Confidence interval: 1,127.60 – 1,145.21
7446 votes
Category
hard_prompts
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / hard_prompts_englishhuman preference · 2026-10-28 · rating
Reka Flash 21B (20240226)
1,144.14 rating
Confidence interval: 1,133.69 – 1,154.59
4796 votes
Category
hard_prompts_english
Evaluated configuration
Reka Flash 21B (20240226)
Evaluation type
human preference
Arena text preference / industry_business_and_management_and_financial_operationshuman preference · 2026-10-28 · rating