Scores updated · 21 tests both models report · How we compare
Every test, side by side
All 21 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Other results21 tests
- SimpleQAHy4-flash-base by 1126.537.5+11
- U-MATHHy4-flash-base by 8.960.969.8+8.9
- SuperGPQAHy4-flash-base by 6.251.657.8+6.2
- MMLU-ProHy4-flash-base by 5.365.871.1+5.3
- MBPP+ (EvalPlus-augmented)Hy3-preview-Base by 578.773.7+5
- Chinese SimpleQA (C-SimpleQA)Hy4-flash-base by 4.969.774.6+4.9
- IncludeHy4-flash-base by 4.278.682.8+4.2
- AGIEval (Chinese)Hy4-flash-base by 2.887.890.6+2.8
- C-EvalHy4-flash-base by 2.289.892+2.2
- CMMLUHy4-flash-base by 289.691.6+2
- C3Hy4-flash-base by 1.475.977.3+1.4
- OCW CoursesHy4-flash-base by 1.462.664+1.4
- MMLUHy4-flash-base by 1.387.488.7+1.3
- ARC-ChallengeHy3-preview-Base by 19695+1
- MMLU-Reduxtie86.987.6tie
- AGIEval-Entie7675.4tie
- HellaSwagtie85.785.2tie
- Belebele (5-language subset)tie94.293.8tie
- Social IQAtie80.580.1tie
- KOR-Benchtie54.654.7tie
- PIQAtie84.484.4tie
Questions people ask
How do you compare the two?
We use the 21 benchmark tests both models have published scores on. None of them is in the eight capability areas we count, so this page lists them without a verdict. Each score is the one shown on the model's own page.