Claude Sonnet 5 (Adaptive Reasoning, High Effort): benchmark results
Provider: Anthropic. Released 2026-06-30. Access: API.
Unified ELO 1622 ± 1, rank #365 of 1919 rated models, from 16 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Writing | 56.79 | Writing Score | 91 |
| UGI - Natural Intelligence | 43.75 | NatInt Score | 87.7 |
| AA Humanity's Last Exam | 35.73 | Accuracy (%) | 87.4 |
| AA Omniscience - Software Engineering (SWE) | 58.6 | Accuracy (%) | 83.2 |
| AA Long Context Reasoning | 76.67 | Accuracy (%) | 80.9 |
| AA-Omniscience Accuracy | 37.43 | Accuracy (%) | 79.9 |
| AA Omniscience - Science, Engineering & Mathematics | 40.8 | Accuracy (%) | 79.6 |
| AA Omniscience - Business | 29.1 | Accuracy (%) | 79.4 |
| AA Omniscience - Humanities & Social Sciences | 34.5 | Accuracy (%) | 79 |
| AA Omniscience - Health | 33.8 | Accuracy (%) | 78.8 |
| AA Omniscience - Law | 27.8 | Accuracy (%) | 78.6 |
| AA Omniscience | -3.68 | Score | 78 |
Interactive version: theaggregate.ai/model?slug=claude-sonnet-5-adaptive-reasoning-high-effort · How It Works · Data refreshed daily, snapshot 2026-09-08.