GPT-5 vs Claude Opus 4.7, which Frontier Model Wins in 2026?

GPT-5Claude Opus 4.7
arenaElo14121420
humanEval91.693.1
sweBench6564.3
mmlu91.592.3
gpqa63.865.2
aime9288
Price in/out ($/M)$1.25 / $10$5 / $25

Our verdict

Claude Opus 4.7 leads on long-context coding and agentic workflows; GPT-5 wins on multimodal breadth (native audio) and tooling maturity. For enterprise coding agents, pick Opus 4.7. For voice-first multimodal apps, pick GPT-5.

GPT-5 · Claude Opus 4.7 · All AI model comparisons