← All comparisons
Claude 4.1 Opus (Reasoning) vs DeepSeek V3.2 Exp (Reasoning)
Anthropic vs DeepSeek — side-by-side benchmark comparison
| Claude 4.1 Opus (Reasoning) | DeepSeek V3.2 Exp (Reasoning) | |
|---|---|---|
| Intelligence Index | 42.0 | 32.9 |
| Coding Index | 36.5 | 33.3 |
| Math Index | 80.3 | 87.7 |
| Output speed (tok/s) | 44.5 | 0.0 |
| Blended price ($/1M) | $32.81 | $0.31 |
| Time to first token (s) | 8.55s | 0.00s |
| aime | — | — |
| aime 25 | 80.3% | 87.7% |
| artificial analysis coding index | 36.50 | 33.30 |
| artificial analysis intelligence index | 42.00 | 32.90 |
| artificial analysis math index | 80.30 | 87.70 |
| gpqa | 80.9% | 79.7% |
| hle | 11.9% | 13.8% |
| ifbench | 55.4% | 54.1% |
| lcr | 66.3% | 69.0% |
| livecodebench | 65.4% | 78.9% |
| math 500 | — | — |
| mmlu pro | 88.0% | 85.0% |
| scicode | 40.9% | 37.7% |
| tau2 | 71.4% | 33.9% |
| terminalbench hard | 34.3% | 31.1% |
Benchmark data from Artificial Analysis.