Claude Sonnet 5 (Adaptive Reasoning, Xhigh Effort): benchmark results
Provider: Anthropic. Released 2026-06-30. Access: API.
Unified ELO 1631 ± 1, rank #333 of 1919 rated models, from 16 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Writing | 58.45 | Writing Score | 91.9 |
| AA Humanity's Last Exam | 39.02 | Accuracy (%) | 90.3 |
| UGI - Natural Intelligence | 48.36 | NatInt Score | 89.5 |
| AA Omniscience - Science, Engineering & Mathematics | 44.2 | Accuracy (%) | 85.1 |
| AA Omniscience | 3.18 | Score | 84.2 |
| AA Omniscience - Software Engineering (SWE) | 59 | Accuracy (%) | 83.4 |
| AA Omniscience - Business | 32.3 | Accuracy (%) | 82.1 |
| AA-Omniscience Accuracy | 38.97 | Accuracy (%) | 82 |
| AA Omniscience - Humanities & Social Sciences | 35.9 | Accuracy (%) | 81.2 |
| AA Long Context Reasoning | 76.67 | Accuracy (%) | 80.9 |
| UGI Leaderboard | 45.26 | UGI Score | 80.6 |
| AA Omniscience - Health | 34.1 | Accuracy (%) | 79.2 |
Interactive version: theaggregate.ai/model?slug=claude-sonnet-5-adaptive-reasoning-xhigh-effort · How It Works · Data refreshed daily, snapshot 2026-09-08.