← All comparisons
Magistral Small 1.2 vs Hermes 4 - Llama-3.1 70B (Reasoning)
Mistral vs Nous Research — side-by-side benchmark comparison
| Magistral Small 1.2 | Hermes 4 - Llama-3.1 70B (Reasoning) | |
|---|---|---|
| Intelligence Index | 18.2 | 16.0 |
| Coding Index | 14.8 | 14.4 |
| Math Index | 80.3 | 68.7 |
| Output speed (tok/s) | 111.1 | 92.8 |
| Blended price ($/1M) | $0.75 | $0.20 |
| Time to first token (s) | 0.38s | 0.64s |
| aime | — | — |
| aime 25 | 80.3% | 68.7% |
| artificial analysis coding index | 14.80 | 14.40 |
| artificial analysis intelligence index | 18.20 | 16.00 |
| artificial analysis math index | 80.30 | 68.70 |
| gpqa | 66.3% | 69.9% |
| hle | 6.1% | 7.9% |
| ifbench | 44.4% | 31.3% |
| lcr | 16.3% | 6.7% |
| livecodebench | 72.3% | 65.3% |
| math 500 | — | — |
| mmlu pro | 76.8% | 81.1% |
| scicode | 35.2% | 34.1% |
| tau2 | 27.8% | 22.5% |
| terminalbench hard | 4.5% | 4.5% |
Benchmark data from Artificial Analysis.