Llama 3.1 Instruct 405B is the stronger all-rounder.
Scores updated · 15 tests both models report · How we compare
Where each one wins
Tests won in each of the one area where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
Reasoning rests on a single test.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Llama 3.1 Instruct 405B pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+7.6points ahead
Where Mistral Large 3 675B Base 2512 pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 15 tests both models report. The winning score is in its model's colour; marks a score checked independently.
ReasoningLlama 3.1 Instruct 405B
- GPQA DiamondLlama 3.1 Instruct 405B by 7.651.543.9+7.6
Other results14 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- OpenBookQALlama 3.1 Instruct 405B by 42.69451.4+42.6
- HumanEvalLlama 3.1 Instruct 405B by 22.38966.7+22.3
- MMMLUMistral Large 3 675B Base 2512 by 11.773.885.5+11.7
- MATHLlama 3.1 Instruct 405B by 10.973.862.9+10.9
- MGSMLlama 3.1 Instruct 405B by 8.791.682.9+8.7
- LiveCodeBenchMistral Large 3 675B Base 2512 by 6.727.734.4+6.7
- SimpleQAMistral Large 3 675B Base 2512 by 6.717.123.8+6.7
- MMLU-ProLlama 3.1 Instruct 405B by 5.973.367.4+5.9
- GSM8KLlama 3.1 Instruct 405B by 5.696.891.2+5.6
- WinoGrandeLlama 3.1 Instruct 405B by 3.185.282.1+3.1
- MMLULlama 3.1 Instruct 405B by 1.288.687.3+1.2
- PIQALlama 3.1 Instruct 405B by 1.185.984.8+1.1
- ARC-Challengetie96.997.3tie
- HellaSwagtie89.288.9tie
Questions people ask
Which is better, Llama 3.1 Instruct 405B or Mistral Large 3 675B Base 2512?
Llama 3.1 Instruct 405B wins the one area where both have results: reasoning. Mistral Large 3 675B Base 2512 wins none.
How do you compare the two?
We use the 15 benchmark tests both models have published scores on. The verdict counts the 1 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 14 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.