← All comparisons
Magistral Medium 1.2 vs Qwen3 32B (Reasoning)
Mistral vs Alibaba — side-by-side benchmark comparison
| Magistral Medium 1.2 | Qwen3 32B (Reasoning) | |
|---|---|---|
| Intelligence Index | 27.1 | 16.5 |
| Coding Index | 21.7 | 13.8 |
| Math Index | 82.0 | 73.0 |
| Output speed (tok/s) | 39.4 | 93.4 |
| Blended price ($/1M) | $2.75 | $0.28 |
| Time to first token (s) | 0.54s | 1.13s |
| aime | — | 80.7% |
| aime 25 | 82.0% | 73.0% |
| artificial analysis coding index | 21.70 | 13.80 |
| artificial analysis intelligence index | 27.10 | 16.50 |
| artificial analysis math index | 82.00 | 73.00 |
| gpqa | 73.9% | 66.8% |
| hle | 9.6% | 8.3% |
| ifbench | 43.0% | 36.3% |
| lcr | 51.3% | 0.0% |
| livecodebench | 75.0% | 54.6% |
| math 500 | — | 96.1% |
| mmlu pro | 81.5% | 79.8% |
| scicode | 39.2% | 35.4% |
| tau2 | 52.0% | 29.8% |
| terminalbench hard | 12.9% | 3.0% |
Benchmark data from Artificial Analysis.