Scores updated · 11 tests both models report · How we compare
Every test, side by side
All 11 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Other results11 tests
- simple_safety_testsMistral Instruct V0.3 (7B) by 27.553.581+27.5
- NarrativeQAMistral Instruct V0.3 (7B) by 22.848.871.6+22.8
- XSTestMistral Instruct V0.3 (7B) by 15.277.492.6+15.2
- anthropic_red_teamMistral Instruct V0.3 (7B) by 13.576.690.1+13.5
- GSM8KDBRX Instruct by 13.367.153.8+13.3
- MMLUDBRX Instruct by 13.364.351+13.3
- bbqDBRX Instruct by 12.479.266.8+12.4
- OpenBookQADBRX Instruct by 129179+12
- HarmBenchMistral Instruct V0.3 (7B) by 7.327.134.4+7.3
- AIR-Bench 2024Mistral Instruct V0.3 (7B) by 7.225.432.5+7.2
- naturalquestions_closedbookDBRX Instruct by 3.128.425.3+3.1
Questions people ask
How do you compare the two?
We use the 11 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.