Scores updated · 13 tests both models report · How we compare
Every test, side by side
All 13 tests both models report. The winning score is in its model's colour; marks a score checked independently.
Other results13 tests
- HumanEvalLlama 3.1 Instruct 405B by 48.840.289+48.8
- ARC-ChallengeLlama 3.1 Instruct 405B by 28.568.496.9+28.5
- GSM8KLlama 3.1 Instruct 405B by 28.268.696.8+28.2
- MMLULlama 3.1 Instruct 405B by 24.164.588.6+24.1
- MBPPLlama 3.1 Instruct 405B by 1652.468.4+16
- naturalquestions_closedbookLlama 3.1 Instruct 405B by 12.832.845.6+12.8
- NaturalQuestionsLlama 3.1 Instruct 405B by 12.329.241.5+12.3
- HellaSwagLlama 3.1 Instruct 405B by 7.381.989.2+7.3
- WinoGrandeLlama 3.1 Instruct 405B by 4.680.685.2+4.6
- PIQALlama 3.1 Instruct 405B by 4.281.785.9+4.2
- OpenBookQALlama 3.1 Instruct 405B by 39194+3
- NarrativeQAGemma 2 Instruct (9B) by 1.976.874.9+1.9
- MATHLlama 3.1 Instruct 405B by 1.472.473.8+1.4
Questions people ask
How do you compare the two?
We use the 13 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.