Scores updated · 20 tests both models report · How we compare
Every test, side by side
All 20 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Other results20 tests
- MultiPL-E HumanEvalStep3.5 Flash Base by 2245.767.7+22
- HumanEvalStep3.5 Flash Base by 19.261.981.1+19.2
- MultiPL-E MBPPStep3.5 Flash Base by 7.450.658+7.4
- MATHStep3.5 Flash Base by 6.760.166.8+6.7
- Chinese SimpleQA (C-SimpleQA)DeepSeek-V3.2-Exp-Base by 4.86863.2+4.8
- SimpleQAStep3.5 Flash Base by 4.62731.6+4.6
- GPQA (unspecified)Step3.5 Flash Base by 4.437.341.7+4.4
- HumanEval+Step3.5 Flash Base by 4.367.772+4.3
- WinoGrandeDeepSeek-V3.2-Exp-Base by 4.383.479.1+4.3
- GSM8KStep3.5 Flash Base by 3.884.488.2+3.8
- MBPPStep3.5 Flash Base by 3.875.679.4+3.8
- SuperGPQADeepSeek-V3.2-Exp-Base by 2.643.641+2.6
- MMLUDeepSeek-V3.2-Exp-Base by 287.885.8+2
- C-EvalDeepSeek-V3.2-Exp-Base by 1.49189.6+1.4
- MMLU-ReduxDeepSeek-V3.2-Exp-Base by 1.290.489.2+1.2
- MMLU-ProDeepSeek-V3.2-Exp-Base by 163.362.3+1
- HellaSwagtie89.490.2tie
- MBPP+ (EvalPlus-augmented)tie69.870.6tie
- BBHtie88.788.2tie
- CMMLUtie88.988.9tie
Questions people ask
How do you compare the two?
We use the 20 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.