Interval Quality on SWE-bench
44.3Hit RateGPT-5.2 Instant
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT-5.2 Instant2026.05 | 44.3 | 11.5 | |
| Claude Opus 4.72026.05 | 30.3 | 16 | |
| Gemini 3.1 Pro Preview2026.05 | 23.2 | 11.2 | |
| Claude Sonnet 4.62026.05 | 22.3 | 13 | |
| Qwen3 235B2026.05 | 6.5 | 2.1 |