← All comparisons
GPT-5.1 Codex (high) vs Claude 3.5 Sonnet (June '24)
OpenAI vs Anthropic — side-by-side benchmark comparison
| GPT-5.1 Codex (high) | Claude 3.5 Sonnet (June '24) | |
|---|---|---|
| Intelligence Index | 43.1 | 14.2 |
| Coding Index | 36.6 | 26.0 |
| Math Index | 95.7 | — |
| Output speed (tok/s) | 182.1 | 0.0 |
| Blended price ($/1M) | $3.44 | $6.56 |
| Time to first token (s) | 5.42s | 0.00s |
| aime | — | 9.7% |
| aime 25 | 95.7% | — |
| artificial analysis coding index | 36.60 | 26.00 |
| artificial analysis intelligence index | 43.10 | 14.20 |
| artificial analysis math index | 95.70 | — |
| gpqa | 86.0% | 56.0% |
| hle | 23.4% | 3.7% |
| ifbench | 70.0% | — |
| lcr | 67.3% | — |
| livecodebench | 84.9% | — |
| math 500 | — | 69.5% |
| mmlu pro | 86.0% | 75.1% |
| scicode | 40.2% | 31.6% |
| tau2 | 83.0% | — |
| terminalbench hard | 34.8% | — |
Benchmark data from Artificial Analysis.