Feasibility Prediction on SWE-bench
47.6F1@1Qwen3 235B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3 235B2026.05 | 47.6 | 35.1 | 32.8 | |
| GPT-5.2 Instant2026.05 | 43.5 | 40.2 | 21.2 | |
| Claude Opus 4.72026.05 | 41.1 | 51.1 | 48.8 | |
| Gemini 3.1 Pro Preview2026.05 | 39.2 | 58.2 | 52 | |
| Claude Sonnet 4.62026.05 | 32.2 | 37.7 | 23.4 |