Claude Opus 4.7 (Adaptive Reasoning, Max Effort): benchmark results

Claude Opus 4.7 evaluated in adaptive-reasoning mode at max effort. Provider: Anthropic. Released 2026-04-16. Access: API.

Unified ELO 1667 ± 1, rank #193 of 1761 rated models, from 30 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
ITBench-AA46.7Average Precision at Full Recall (self-reported)95.5
AA Terminal-Bench Hard51.52Accuracy (%)95.4
AA Omniscience27.27Score95.3
UGI - Natural Intelligence62.94NatInt Score95
AA Omniscience - Science, Engineering & Mathematics51.77Accuracy (%)94.3
AA Omniscience - Software Engineering (SWE)75.3Accuracy (%)93.8
UGI - Writing61.4Writing Score93.6
Artificial Analysis Intelligence Index44.29Intelligence Index93.5
AA Humanity's Last Exam42.31Accuracy (%)93.3
UGI Leaderboard51.79UGI Score92.9
AA GPQA Diamond91.41Accuracy (%)92.5
AA Omniscience - Business42.25Accuracy (%)92.2

Interactive version: theaggregate.ai/model?slug=claude-opus-4-7-adaptive-reasoning-max-effort · How It Works · Data refreshed daily, snapshot 2026-09-05.