Claude Sonnet 5 (Adaptive Reasoning, Low Effort): benchmark results
Provider: Anthropic. Released 2026-06-30. Access: API.
Unified ELO 1598 ± 1, rank #470 of 1919 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Writing | 53 | Writing Score | 89.4 |
| UGI - Natural Intelligence | 40.31 | NatInt Score | 86.5 |
| AA Omniscience - Software Engineering (SWE) | 55.7 | Accuracy (%) | 81.9 |
| AA Omniscience - Business | 30.8 | Accuracy (%) | 80.3 |
| Artificial Analysis Intelligence Index | 24.74 | Intelligence Index | 80 |
| AA Omniscience - Health | 34.4 | Accuracy (%) | 79.5 |
| AA-Omniscience Accuracy | 37.35 | Accuracy (%) | 79.5 |
| AA Omniscience - Humanities & Social Sciences | 34.6 | Accuracy (%) | 79.2 |
| AA Omniscience - Law | 28.1 | Accuracy (%) | 78.8 |
| AA Omniscience - Science, Engineering & Mathematics | 40.5 | Accuracy (%) | 78.1 |
| AA CritPt | 4.57 | Accuracy (%) | 77.1 |
| AA Omniscience | -8.23 | Score | 74.4 |
Interactive version: theaggregate.ai/model?slug=claude-sonnet-5-adaptive-reasoning-low-effort · How It Works · Data refreshed daily, snapshot 2026-09-08.