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 Midtrain by 17.345.763+17.3
- Chinese SimpleQA (C-SimpleQA)DeepSeek-V3.2-Exp-Base by 9.96858.1+9.9
- WinoGrandeDeepSeek-V3.2-Exp-Base by 9.885.675.8+9.8
- MBPP+ (EvalPlus-augmented)DeepSeek-V3.2-Exp-Base by 7.469.862.4+7.4
- HumanEval+Step3.5 Flash Base Midtrain by 7.367.775+7.3
- GPQA (unspecified)Step3.5 Flash Base Midtrain by 6.537.343.8+6.5
- HumanEvalStep3.5 Flash Base Midtrain by 5.161.967+5.1
- MMLUDeepSeek-V3.2-Exp-Base by 4.487.883.4+4.4
- C-EvalDeepSeek-V3.2-Exp-Base by 3.89187.2+3.8
- MBPPStep3.5 Flash Base Midtrain by 3.475.679+3.4
- MATHStep3.5 Flash Base Midtrain by 3.262.565.7+3.2
- MMLU-ReduxDeepSeek-V3.2-Exp-Base by 3.290.487.2+3.2
- MultiPL-E MBPPDeepSeek-V3.2-Exp-Base by 2.750.647.9+2.7
- SuperGPQADeepSeek-V3.2-Exp-Base by 2.443.641.2+2.4
- GSM8KDeepSeek-V3.2-Exp-Base by 2.291.188.9+2.2
- CMMLUDeepSeek-V3.2-Exp-Base by 288.986.9+2
- HellaSwagStep3.5 Flash Base Midtrain by 1.689.491+1.6
- MMLU-ProStep3.5 Flash Base Midtrain by 1.562.163.6+1.5
- BBHDeepSeek-V3.2-Exp-Base by 1.488.787.3+1.4
- SimpleQAStep3.5 Flash Base Midtrain by 1.42728.4+1.4
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.