Sonnet 4.6 is Anthropic's most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work. It excels at iterative development, complex codebase navigation, end-to-end project management with...
| Benchmark | Score | Source |
|---|---|---|
| MMMLU | 89.3 | verified |
| Benchmark | Score | Source |
|---|---|---|
| SWE-bench Verified | 79.6 | verified |
| Benchmark | Score | Source |
|---|
| GPQA Diamond | 89.9 | verified |
| ARC-AGI v2 | 58.3 | verified |
| Humanity's Last Exam | 49 | verified |
| Benchmark | Score | Source |
|---|---|---|
| MMMU-Pro | 75.6 | verified |
| Benchmark | Score | Source |
|---|---|---|
| BrowseComp | 74.7 | verified |
| OSWorld | 72.5 | verified |
| MCP Atlas | 61.3 | verified |
Anthropic