← All comparisons
Claude Sonnet 4.6 (Non-reasoning, Low Effort) vs DeepSeek V4 Pro (Non-reasoning)
Anthropic vs DeepSeek — side-by-side benchmark comparison
| Claude Sonnet 4.6 (Non-reasoning, Low Effort) | DeepSeek V4 Pro (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 42.6 | 39.3 |
| Coding Index | 43.0 | 38.4 |
| Math Index | — | — |
| Output speed (tok/s) | 54.9 | 53.3 |
| Blended price ($/1M) | $6.56 | $0.54 |
| Time to first token (s) | 1.13s | 1.18s |
| aime | — | — |
| aime 25 | — | — |
| artificial analysis coding index | 43.00 | 38.40 |
| artificial analysis intelligence index | 42.60 | 39.30 |
| artificial analysis math index | — | — |
| gpqa | 79.7% | 71.7% |
| hle | 10.8% | 7.7% |
| ifbench | 42.4% | 45.8% |
| lcr | 58.7% | 45.0% |
| livecodebench | — | — |
| math 500 | — | — |
| mmlu pro | — | — |
| scicode | 44.1% | 42.4% |
| tau2 | 78.9% | 91.2% |
| terminalbench hard | 42.4% | 36.4% |
Benchmark data from Artificial Analysis.