Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max,...
| Benchmark | Score | Source |
|---|---|---|
| SWE-bench Verified | 85.2 | verified |
| SWE-Bench Pro | 63.2 | verified |
| Benchmark | Score | Source |
|---|---|---|
| Humanity's Last Exam | 57.4 | verified |
| Benchmark | Score | Source |
|---|---|---|
| CharXiv Reasoning | 88.3 | verified |
| Benchmark | Score | Source |
|---|---|---|
| BrowseComp | 84.7 | verified |
| Toolathlon | 54.3 | verified |
Anthropic