Claude Sonnet 4 (High): benchmark results
Provider: Anthropic. Released 2025-05-22. Access: API.
Unified ELO 1633 ± 1, rank #430 of 3078 rated models, from 12 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| IKP - T2 | 100 | Accuracy (%) | 92.6 |
| HAL Online Mind2Web | 39.33 | Accuracy (%) | 77.3 |
| MCPMark | 28.35 | Pass@1 (%) | 65.8 |
| ObviousBench | 95.83 | Answer pass³ (%) | 65.8 |
| IKP - T1 | 99.5 | Accuracy (%) | 58.5 |
| HAL CORE-Bench Hard | 33.33 | Accuracy (%) | 48.1 |
| IKP - T4 | 70.21 | Accuracy (%) | 43.1 |
| Incompressible Knowledge Probes (IKP) | 57.41 | Mean tier accuracy (%) | 42.4 |
| IKP - T5 | 33.7 | Accuracy (%) | 41.6 |
| IKP - T3 | 88.52 | Accuracy (%) | 39.5 |
| IKP - T6 | 6.42 | Accuracy (%) | 28.9 |
| IKP - T7 | 3.49 | Accuracy (%) | 21.3 |
Interactive version: theaggregate.ai/model?slug=claude-sonnet-4-high · How It Works · Data refreshed daily, snapshot 2026-09-19.