Scores updated · 14 tests both models report · How we compare
Every test, side by side
All 14 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Other results14 tests
- Arena HardGranite 3.2 8B Instruct by 26.455.328.9+26.4
- AlpacaEval-2.0Granite 3.2 8B Instruct by 17.761.243.5+17.7
- DROPGranite 3.2 8B Instruct by 1458.344.3+14
- MMLUGranite 3.2 8B Instruct by 10.966.855.9+10.9
- HumanEval+Granite 3.2 8B Instruct by 1085.775.7+10
- PopQAGranite 3.2 8B Instruct by 9.62818.4+9.6
- GSM8KGranite 3.2 8B Instruct by 9.281.772.5+9.2
- HumanEvalGranite 3.2 8B Instruct by 8.989.380.5+8.9
- IFEvalGranite 3.2 8B Instruct by 8.574.365.8+8.5
- BigBenchHardGranite 3.2 8B Instruct by 871.963.9+8
- TruthfulQAGranite 3.2 8B Instruct by 7.966.959+7.9
- MATH-500 (EM)Granite 3.3 2B Instruct by 5.352.858.1+5.3
- AttaQGranite 3.3 2B Instruct by 2.884.787.5+2.8
- AIME 2024tie2.43.3tie
Questions people ask
How do you compare the two?
We use the 14 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.