← All comparisons
Magistral Small 1.2 vs Llama Nemotron Super 49B v1.5 (Reasoning)
Mistral vs NVIDIA — side-by-side benchmark comparison
| Magistral Small 1.2 | Llama Nemotron Super 49B v1.5 (Reasoning) | |
|---|---|---|
| Intelligence Index | 18.2 | 18.7 |
| Coding Index | 14.8 | 15.1 |
| Math Index | 80.3 | 76.7 |
| Output speed (tok/s) | 111.1 | 48.1 |
| Blended price ($/1M) | $0.75 | $0.17 |
| Time to first token (s) | 0.38s | 0.30s |
| aime | — | 86.0% |
| aime 25 | 80.3% | 76.7% |
| artificial analysis coding index | 14.80 | 15.10 |
| artificial analysis intelligence index | 18.20 | 18.70 |
| artificial analysis math index | 80.30 | 76.70 |
| gpqa | 66.3% | 74.8% |
| hle | 6.1% | 6.8% |
| ifbench | 44.4% | 37.0% |
| lcr | 16.3% | 34.0% |
| livecodebench | 72.3% | 73.7% |
| math 500 | — | 98.3% |
| mmlu pro | 76.8% | 81.4% |
| scicode | 35.2% | 34.8% |
| tau2 | 27.8% | 28.1% |
| terminalbench hard | 4.5% | 5.3% |
Benchmark data from Artificial Analysis.