← All comparisons
GPT-5.3 Codex (xhigh) vs Claude 4.1 Opus (Non-reasoning)
OpenAI vs Anthropic — side-by-side benchmark comparison
| GPT-5.3 Codex (xhigh) | Claude 4.1 Opus (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 53.6 | 36.0 |
| Coding Index | 53.1 | — |
| Math Index | — | — |
| Output speed (tok/s) | 84.7 | 44.7 |
| Blended price ($/1M) | $4.81 | $32.81 |
| Time to first token (s) | 60.27s | 1.63s |
| aime | — | — |
| aime 25 | — | — |
| artificial analysis coding index | 53.10 | — |
| artificial analysis intelligence index | 53.60 | 36.00 |
| artificial analysis math index | — | — |
| gpqa | 91.5% | — |
| hle | 39.9% | — |
| ifbench | 75.4% | — |
| lcr | 74.0% | — |
| livecodebench | — | — |
| math 500 | — | — |
| mmlu pro | — | — |
| scicode | 53.2% | — |
| tau2 | 86.0% | — |
| terminalbench hard | 53.0% | — |
Benchmark data from Artificial Analysis.