← All comparisons
Ministral 3 8B vs Claude Opus 4.5 (Reasoning)
Mistral vs Anthropic — side-by-side benchmark comparison
| Ministral 3 8B | Claude Opus 4.5 (Reasoning) | |
|---|---|---|
| Intelligence Index | 14.8 | 49.7 |
| Coding Index | 10.0 | 47.8 |
| Math Index | 31.7 | 91.3 |
| Output speed (tok/s) | 100.9 | 73.2 |
| Blended price ($/1M) | $0.15 | $10.94 |
| Time to first token (s) | 0.37s | 11.69s |
| aime | — | — |
| aime 25 | 31.7% | 91.3% |
| artificial analysis coding index | 10.00 | 47.80 |
| artificial analysis intelligence index | 14.80 | 49.70 |
| artificial analysis math index | 31.70 | 91.30 |
| gpqa | 47.1% | 86.6% |
| hle | 4.3% | 28.4% |
| ifbench | 29.1% | 58.0% |
| lcr | 24.0% | 74.0% |
| livecodebench | 30.3% | 87.1% |
| math 500 | — | — |
| mmlu pro | 64.2% | 89.5% |
| scicode | 20.8% | 49.5% |
| tau2 | 26.6% | 89.5% |
| terminalbench hard | 4.5% | 47.0% |
Benchmark data from Artificial Analysis.