Calibration
SWE-bench Pro → CursorBench 3.2
CursorBench 3.2 is estimated from SWE-bench Pro with a Hill curve fitted on 12 models measured on both: y = 0.0000 + (0.7496 − 0.0000)·x^6.00 / (0.48799^6.00 + x^6.00), R² = 0.81, cross-validated error 3.6 pp. It is used for 55 estimates.
| Estimated model | SWE-bench Pro | CursorBench 3.2 | Source |
|---|---|---|---|
| Atria Dawn Preview | 59.6% | 57.6% | estimated ± 3.6 pp, high confidence |
| Claude Mythos 5 | 80.3% | 71.4% | estimated ± 3.6 pp, high confidence |
| Claude Opus 4.5 | 57.1% | 53.9% | estimated ± 3.6 pp, high confidence |
| Claude Opus 4.7 (Adaptive) | 64.3% | 62.9% | estimated ± 3.6 pp, high confidence |
| DeepSeek V4 Pro 0813 | 55.4% | 51.1% | estimated ± 3.6 pp, high confidence |
| dots3-note Preview | 61.0% | 59.4% | estimated ± 3.6 pp, high confidence |
| Gemini 3.5 Flash-Lite | 54.2% | 48.9% | estimated ± 3.6 pp, medium confidence |
| GLM-5 | 55.1% | 50.6% | estimated ± 3.6 pp, high confidence |
| GLM-5.1 | 58.4% | 55.9% | estimated ± 3.6 pp, high confidence |
| GPT-5.2 | 55.6% | 51.4% | estimated ± 3.6 pp, high confidence |
| GPT-5.3 Codex | 56.8% | 53.5% | estimated ± 3.6 pp, high confidence |
| GPT-5.4 | 57.7% | 54.9% | estimated ± 3.6 pp, high confidence |
| Granite 4.2 30B | 33.3% | 6.9% | estimated ± 3.6 pp, medium confidence |
| Granite 4.2 8B | 19.1% | 0.3% | estimated ± 3.6 pp, medium confidence |
| Grok 4.20 | 51.8% | 44.1% | estimated ± 3.6 pp, medium confidence |
| Hy4 preview | 65.7% | 64.2% | estimated ± 3.6 pp, high confidence |
| Inkling | 54.3% | 49.1% | estimated ± 3.6 pp, medium confidence |
| Inkling-Small | 55.9% | 52.0% | estimated ± 3.6 pp, high confidence |
| Kimi K2.6 | 58.6% | 56.2% | estimated ± 3.6 pp, high confidence |
| Kimi K2.5 | 50.7% | 41.8% | estimated ± 3.6 pp, medium confidence |
| Laguna M.1 | 49.2% | 38.4% | estimated ± 3.6 pp, medium confidence |
| Laguna S 2.1 | 59.4% | 57.3% | estimated ± 3.6 pp, high confidence |
| Laguna XS.2 | 46.3% | 31.6% | estimated ± 3.6 pp, medium confidence |
| Laguna XS 2.1 | 47.6% | 34.7% | estimated ± 3.6 pp, medium confidence |
| Ling 3.0 Flash | 56.6% | 53.1% | estimated ± 3.6 pp, high confidence |
| LLaDA2.2-flash | 30.1% | 3.9% | estimated ± 3.6 pp, medium confidence |
| LongCat-Flash-Lite-Sparse | 40.6% | 18.7% | estimated ± 3.6 pp, medium confidence |
| MAI-Thinking-1 | 52.8% | 46.2% | estimated ± 3.6 pp, medium confidence |
| MiMo-V2.5 | 56.1% | 52.3% | estimated ± 3.6 pp, high confidence |
| MiMo-V2.5-Pro | 57.2% | 54.1% | estimated ± 3.6 pp, high confidence |
| MiniCPM5-2B | 14.4% | 0.0% | estimated ± 3.6 pp, medium confidence |
| MiniMax M2.7 | 56.2% | 52.5% | estimated ± 3.6 pp, high confidence |
| MiniMax M3 | 59.0% | 56.8% | estimated ± 3.6 pp, high confidence |
| Muse Glimmer 30B | 51.2% | 42.8% | estimated ± 3.6 pp, medium confidence |
| Muse Spark | 52.4% | 45.4% | estimated ± 3.6 pp, medium confidence |
| Muse Spark 1.1 | 61.5% | 60.0% | estimated ± 3.6 pp, high confidence |
| Ornith-1.0-35B | 50.4% | 41.1% | estimated ± 3.6 pp, medium confidence |
| Ornith-1.0-397B | 62.2% | 60.8% | estimated ± 3.6 pp, high confidence |
| Ornith-1.0-9B | 42.9% | 23.7% | estimated ± 3.6 pp, medium confidence |
| Ornith-1.5-35B-A3B | 59.6% | 57.6% | estimated ± 3.6 pp, high confidence |
| Ornith-1.5-397B | 65.1% | 63.7% | estimated ± 3.6 pp, high confidence |
| Ornith-1.5-9B | 47.5% | 34.5% | estimated ± 3.6 pp, medium confidence |
| Qwen3.5 397B | 50.9% | 42.2% | estimated ± 3.6 pp, medium confidence |
| Qwen3.6-27B | 53.5% | 47.6% | estimated ± 3.6 pp, medium confidence |
| Qwen3.6-35B-A3B | 49.5% | 39.1% | estimated ± 3.6 pp, medium confidence |
| Qwen 3.6 Max (preview) | 57.3% | 54.3% | estimated ± 3.6 pp, high confidence |
| Qwen3.6 Plus | 56.6% | 53.1% | estimated ± 3.6 pp, high confidence |
| Qwen3.7 Max | 60.6% | 58.9% | estimated ± 3.6 pp, high confidence |
| Qwen3.7 Plus | 57.6% | 54.7% | estimated ± 3.6 pp, high confidence |
| Qwen3.8-Flash-Next | 62.5% | 61.1% | estimated ± 3.6 pp, high confidence |
| Qwen3.8-Omni-Flash | 63.3% | 62.0% | estimated ± 3.6 pp, high confidence |
| Beam | 65.5% | 64.0% | estimated ± 3.6 pp, high confidence |
| Sakana Fugu | 59.0% | 56.8% | estimated ± 3.6 pp, high confidence |
| Sakana Fugu-Ultra | 73.7% | 69.1% | estimated ± 3.6 pp, high confidence |
| Step 3.7 Flash | 56.3% | 52.6% | estimated ± 3.6 pp, high confidence |