Calibration
AA-SciCode → cursorBench41
cursorBench41 is estimated from AA-SciCode with a offset logistic curve fitted on 9 models measured on both: y = 0.4480 + (0.5937 − 0.4480) / (1 + exp(−200.00·(x − 0.6078))), R² = 0.86, cross-validated error 3.7 pp. It is used for 61 estimates.
| Estimated model | AA-SciCode | cursorBench41 | Source |
|---|---|---|---|
| Muse Wonder Full | 62.6% | 59.0% | estimated ± 3.7 pp, high confidence |
| Claude Fable 5 | 61.0% | 53.7% | estimated ± 3.7 pp, high confidence |
| MiMo-V2.6-Pro | 60.9% | 52.9% | estimated ± 3.7 pp, high confidence |
| Kimi K3.1 Pro | 60.8% | 52.2% | estimated ± 3.7 pp, high confidence |
| Kimi K3 | 59.5% | 45.8% | estimated ± 3.7 pp, high confidence |
| Step 5 Preview | 58.9% | 45.1% | estimated ± 3.7 pp, high confidence |
| Muse Spark 1.1 | 58.8% | 45.1% | estimated ± 3.7 pp, high confidence |
| Gemini 3.1 Pro | 58.7% | 45.0% | estimated ± 3.7 pp, high confidence |
| Muse Spark Open | 58.4% | 44.9% | estimated ± 3.7 pp, high confidence |
| GPT-6 Sol | 57.6% | 44.8% | estimated ± 3.7 pp, high confidence |
| Muse Spark 1.2 | 57.4% | 44.8% | estimated ± 3.7 pp, high confidence |
| Gemini 3.7 Flash | 57.2% | 44.8% | estimated ± 3.7 pp, high confidence |
| Gemini 3.8 Flash | 56.6% | 44.8% | estimated ± 3.7 pp, high confidence |
| GPT-6 Astra | 56.5% | 44.8% | estimated ± 3.7 pp, high confidence |
| GPT-5.5 | 55.8% | 44.8% | estimated ± 3.7 pp, high confidence |
| GPT-6.1 Terra | 55.1% | 44.8% | estimated ± 3.7 pp, high confidence |
| Grok 4.5 | 55.0% | 44.8% | estimated ± 3.7 pp, high confidence |
| GPT-6 Luna | 54.6% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Claude Opus 4.8 | 54.4% | 44.8% | estimated ± 3.7 pp, medium confidence |
| GPT-6.1 Sol | 54.2% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Ling 3.1 Flash | 54.1% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Gemini 3.5 Flash | 53.9% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Gemini 3.6 Flash | 53.4% | 44.8% | estimated ± 3.7 pp, medium confidence |
| MiniMax M3.1 Flash | 53.4% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Inkling 2 Small | 52.3% | 44.8% | estimated ± 3.7 pp, medium confidence |
| GPT-5.4 mini | 52.1% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Qwen3.8 Max Preview | 52.1% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Kimi K2.6 | 51.5% | 44.8% | estimated ± 3.7 pp, medium confidence |
| MiMo-V2.6-Flash | 51.3% | 44.8% | estimated ± 3.7 pp, medium confidence |
| GLM-5.2 | 51.2% | 44.8% | estimated ± 3.7 pp, medium confidence |
| DeepSeek V4 Pro | 51.0% | 44.8% | estimated ± 3.7 pp, medium confidence |
| MiMo-V2.5-Pro | 50.6% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Qwen3.8-Flash-Next | 50.6% | 44.8% | estimated ± 3.7 pp, medium confidence |
| DeepSeek V4 Flash 0731 | 50.3% | 44.8% | estimated ± 3.7 pp, medium confidence |
| MiniMax M2.7 | 50.1% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Inkling-Small | 49.7% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Qwen3.7 Max | 49.5% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Hy3 | 48.6% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Hy3 Preview | 48.6% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Beam 1.0 501B-A23B | 48.5% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Grok 4.3 | 48.3% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Quasar 438B | 48.1% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Solar Pro 4 | 48.0% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Kimi K2.7 Code | 47.8% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Nemotron 3.5 Super VL 120B A12B | 47.8% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Nemotron 3.5 Super VL 120B A12B NVFP4 | 47.8% | 44.8% | estimated ± 3.7 pp, medium confidence |
| GPT-5.4 nano | 47.2% | 44.8% | estimated ± 3.7 pp, medium confidence |
| MiniMax M3 | 47.1% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Inkling | 47.0% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Qwen3.8-27B | 46.6% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Gemini 2.5 Pro | 46.3% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Qwen3.7 Plus | 46.1% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Apodex 1.1 | 45.5% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Apodex 1.1 Mini | 45.5% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Gemma 4 31B | 45.5% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Muse Glimmer 30B | 44.9% | 44.8% | estimated ± 3.7 pp, medium confidence |
| GLM-5.1 | 44.8% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Qwen3.5 397B | 44.8% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Ling 3.0 Flash VL | 44.2% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Step 3.7 Flash | 43.9% | 44.8% | estimated ± 3.7 pp, medium confidence |
| Qwen3.6-27B | 42.8% | 44.8% | estimated ± 3.7 pp, medium confidence |
Fitted on
| Model | AA-SciCode | cursorBench41 |
|---|---|---|
| Claude Fable 5.2 | 67.7% | 61.3% |
| Claude Haiku 5.5 | 55.0% | 48.4% |
| Claude Opus 5.5 | 66.9% | 58.1% |
| Claude Sonnet 5.5 | 61.0% | 54.4% |
| Gemini 4 Argon | 61.8% | 56.5% |
| GPT-5.6 Sol | 57.1% | 40.9% |
| Grok 4.6 | 56.5% | 44.3% |
| Grok 4.7 | 57.4% | 48.6% |
| Muse Spark 1.3 | 58.8% | 41.9% |