| GPT-5.4 Codex | Claude Opus 4.7 | |
|---|---|---|
| arenaElo | 1408 | 1420 |
| humanEval | 95.5 | 93.1 |
| sweBench | 70 | 64.3 |
| mmlu | 88.6 | 92.3 |
| gpqa | 60.8 | 65.2 |
| aime | 86 | 88 |
| Price in/out ($/M) | $9 / $36 | $15 / $75 |
5.4 Codex holds the HumanEval crown (95.5 vs 93.1) and is cheaper per token. Opus 4.7 compensates with 2.5× the context and stronger long-horizon agents. Specialized code pipelines: Codex. General coding agents that also reason and plan: Opus.