Calibration
AA-SciCode → SciCode
SciCode is estimated from AA-SciCode with a linear curve fitted on 20 models measured on both: y = 0.9678·x + 0.0131, R² = 0.89, cross-validated error 3.2 pp. It is used for 54 estimates.
| Estimated model | AA-SciCode | SciCode | Source |
|---|---|---|---|
| Apodex 1.1 | 45.5% | 45.3% | estimated ± 3.2 pp, high confidence |
| Apodex 1.1 Mini | 45.5% | 45.3% | estimated ± 3.2 pp, high confidence |
| Celeris-1 | 21.6% | 22.2% | estimated ± 3.2 pp, medium confidence |
| Claude Haiku 5.5 | 55.0% | 54.5% | estimated ± 3.2 pp, high confidence |
| Claude Opus 5.5 | 66.9% | 66.1% | estimated ± 3.2 pp, medium confidence |
| Claude Sonnet 5.5 | 61.0% | 60.3% | estimated ± 3.2 pp, medium confidence |
| Command A+ | 38.5% | 38.6% | estimated ± 3.2 pp, high confidence |
| DeepSeek V3 | 35.8% | 36.0% | estimated ± 3.2 pp, high confidence |
| DeepSeek V3 0324 | 39.0% | 39.1% | estimated ± 3.2 pp, high confidence |
| DeepSeek V4.1 Flash | 51.9% | 51.5% | estimated ± 3.2 pp, high confidence |
| Gemini 2.5 Pro | 46.3% | 46.1% | estimated ± 3.2 pp, high confidence |
| Gemini 4 Argon | 61.8% | 61.1% | estimated ± 3.2 pp, medium confidence |
| Gemma 3 27B | 23.3% | 23.9% | estimated ± 3.2 pp, medium confidence |
| Gemma 4 26B A4B | 40.0% | 40.0% | estimated ± 3.2 pp, high confidence |
| Gemma 4 31B | 45.5% | 45.3% | estimated ± 3.2 pp, high confidence |
| Gemma 4 E4B | 24.4% | 24.9% | estimated ± 3.2 pp, medium confidence |
| GPT-5.6 Luna | 53.6% | 53.2% | estimated ± 3.2 pp, high confidence |
| GPT-6.1 Sol | 54.2% | 53.8% | estimated ± 3.2 pp, high confidence |
| GPT-6 Astra | 56.5% | 56.0% | estimated ± 3.2 pp, high confidence |
| GPT-6 Luna | 54.6% | 54.2% | estimated ± 3.2 pp, high confidence |
| GPT-6 Sol | 57.6% | 57.1% | estimated ± 3.2 pp, high confidence |
| GPT-OSS 120B | 34.0% | 34.2% | estimated ± 3.2 pp, high confidence |
| GPT-OSS 20B | 38.9% | 39.0% | estimated ± 3.2 pp, high confidence |
| Grok 4.7 | 57.4% | 56.9% | estimated ± 3.2 pp, high confidence |
| Hy3 | 48.6% | 48.3% | estimated ± 3.2 pp, high confidence |
| LFM2.5-2.6B | 14.4% | 15.2% | estimated ± 3.2 pp, medium confidence |
| Ling 3.0 Flash VL | 44.2% | 44.1% | estimated ± 3.2 pp, high confidence |
| Ling 3.0 Tiny | 24.2% | 24.7% | estimated ± 3.2 pp, medium confidence |
| Ling 3.1 Flash | 54.1% | 53.7% | estimated ± 3.2 pp, high confidence |
| Llama 4 Maverick | 31.7% | 32.0% | estimated ± 3.2 pp, high confidence |
| Llama 4 Scout | 21.3% | 21.9% | estimated ± 3.2 pp, medium confidence |
| MiMo-V2.6-Flash | 51.3% | 51.0% | estimated ± 3.2 pp, high confidence |
| MiMo-V2.6-Pro | 60.9% | 60.3% | estimated ± 3.2 pp, medium confidence |
| Mistral Large 3 | 36.6% | 36.7% | estimated ± 3.2 pp, high confidence |
| Mistral Large 4 | 54.2% | 53.8% | estimated ± 3.2 pp, high confidence |
| Mistral Medium 3.5 128B | 40.2% | 40.2% | estimated ± 3.2 pp, high confidence |
| Mistral Small 4 | 38.8% | 38.9% | estimated ± 3.2 pp, high confidence |
| Mistral Small 4 (Reasoning) | 38.8% | 38.9% | estimated ± 3.2 pp, high confidence |
| Muse Spark 1.2 | 57.4% | 56.9% | estimated ± 3.2 pp, high confidence |
| Muse Spark 1.3 | 58.8% | 58.2% | estimated ± 3.2 pp, high confidence |
| Nemotron 3 Nano 30B | 30.6% | 30.9% | estimated ± 3.2 pp, high confidence |
| Nemotron 3 Super 100B | 36.2% | 36.3% | estimated ± 3.2 pp, high confidence |
| North Mini Code | 38.8% | 38.9% | estimated ± 3.2 pp, high confidence |
| Quasar 438B | 48.1% | 47.9% | estimated ± 3.2 pp, high confidence |
| Qwen3.5-122B-A10B | 39.7% | 39.7% | estimated ± 3.2 pp, high confidence |
| Qwen3.6-27B | 42.8% | 42.7% | estimated ± 3.2 pp, high confidence |
| Qwen3.6-35B-A3B | 36.6% | 36.7% | estimated ± 3.2 pp, high confidence |
| Qwen3.8-Flash-Next | 50.6% | 50.3% | estimated ± 3.2 pp, high confidence |
| Qwen3.8 Max Preview | 52.1% | 51.7% | estimated ± 3.2 pp, high confidence |
| Solar Pro 3 | 25.5% | 26.0% | estimated ± 3.2 pp, high confidence |
| Solar Pro 4 | 44.6% | 44.5% | estimated ± 3.2 pp, high confidence |
| Step 3.7 Flash | 43.9% | 43.8% | estimated ± 3.2 pp, high confidence |
| Trinity-Large-Preview | 40.6% | 40.6% | estimated ± 3.2 pp, high confidence |
| Trinity-Large-Thinking | 40.6% | 40.6% | estimated ± 3.2 pp, high confidence |