Agentic Coding on SWE-bench-Pro (Resolved)
64.3Resolved RateClaude Opus-4.7
Evaluation Results
| Method | Links | |
|---|---|---|
| Claude Opus-4.72026.06 | 64.3 | |
| OpenAI GPT-5.42026.06 | 59.1 | |
| GLM-5.12026.06 | 58.4 | |
| Gemini 3.1-Pro2026.06 | 54.2 | |
| Ring-2.6-1Tvariant=high2026.06 | 53.76 |
Loading the SOTA2 catalog…