| Claude Opus 4.7 | Claude Sonnet 4.6 | |
|---|---|---|
| arenaElo | 1420 | 1358 |
| humanEval | 93.1 | 89 |
| sweBench | 64.3 | 60 |
| mmlu | 92.3 | 88.7 |
| gpqa | 65.2 | 59.4 |
| aime | 88 | 80 |
| Price in/out ($/M) | $15 / $75 | $3 / $15 |
Opus 4.7 outscores Sonnet 4.6 on every benchmark and offers 5× the context, but costs 5× more per token. For most production chat and coding, Sonnet 4.6 is the sweet spot. Reserve Opus 4.7 for long-context agents, legal analysis and workloads where quality pays for itself.
Claude Opus 4.7 · Claude Sonnet 4.6 · All AI model comparisons