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
- HumanEval+Kimi K2 Base by 12.884.872+12.8
- Chinese SimpleQA (C-SimpleQA)Kimi K2 Base by 11.474.663.2+11.4
- MBPPKimi K2 Base by 9.68979.4+9.6
- MultiPL-E HumanEvalStep3.5 Flash Base by 7.260.567.7+7.2
- MMLU-ProKimi K2 Base by 6.969.262.3+6.9
- WinoGrandeKimi K2 Base by 6.285.379.1+6.2
- HellaSwagKimi K2 Base by 4.494.690.2+4.4
- GSM8KKimi K2 Base by 3.992.188.2+3.9
- SimpleQAKimi K2 Base by 3.735.331.6+3.7
- SuperGPQAKimi K2 Base by 3.744.741+3.7
- MATHKimi K2 Base by 3.470.266.8+3.4
- MBPP+ (EvalPlus-augmented)Kimi K2 Base by 3.273.870.6+3.2
- C-EvalKimi K2 Base by 2.992.589.6+2.9
- HumanEvalStep3.5 Flash Base by 2.978.281.1+2.9
- CMMLUKimi K2 Base by 290.988.9+2
- MMLUKimi K2 Base by 287.885.8+2
- GPQA (unspecified)Kimi K2 Base by 1.443.141.7+1.4
- MMLU-ReduxKimi K2 Base by 190.289.2+1
- MultiPL-E MBPPtie58.858tie
- BBHtie88.788.2tie
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.