← All comparisons
Command A vs Devstral Medium
Cohere vs Mistral — side-by-side benchmark comparison
| Command A | Devstral Medium | |
|---|---|---|
| Intelligence Index | 13.5 | 18.7 |
| Coding Index | 9.9 | 15.9 |
| Math Index | 13.0 | 4.7 |
| Output speed (tok/s) | 53.0 | 77.6 |
| Blended price ($/1M) | $4.38 | $0.80 |
| Time to first token (s) | 0.45s | 0.51s |
| aime | 9.7% | 6.7% |
| aime 25 | 13.0% | 4.7% |
| artificial analysis coding index | 9.90 | 15.90 |
| artificial analysis intelligence index | 13.50 | 18.70 |
| artificial analysis math index | 13.00 | 4.70 |
| gpqa | 52.7% | 49.2% |
| hle | 4.6% | 3.8% |
| ifbench | 36.5% | 29.9% |
| lcr | 18.0% | 28.7% |
| livecodebench | 28.7% | 33.7% |
| math 500 | 81.9% | 70.7% |
| mmlu pro | 71.2% | 70.8% |
| scicode | 28.1% | 29.4% |
| tau2 | 15.2% | 19.9% |
| terminalbench hard | 0.8% | 9.1% |
Benchmark data from Artificial Analysis.