Calibration
GDPval-AA → CWE-bench v1
CWE-bench v1 is estimated from GDPval-AA with a Michaelis–Menten curve fitted on 10 models measured on both: y = 1.3941·x / (0.74812 + x), R² = 0.62, cross-validated error 6.9 pp. It is used for 99 estimates.
| Estimated model | GDPval-AA | CWE-bench v1 | Source |
|---|---|---|---|
| A.X K2 | 23.2% | 33.0% | estimated ± 6.9 pp, low confidence |
| Apodex 1.1 | 34.8% | 44.3% | estimated ± 6.9 pp, medium confidence |
| Apodex 1.1 Mini | 34.8% | 44.3% | estimated ± 6.9 pp, medium confidence |
| Celeris-1 | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Claude Fable 5 | 55.6% | 59.4% | estimated ± 6.9 pp, medium confidence |
| Claude Opus 4.7 (Adaptive) | 42.8% | 50.7% | estimated ± 6.9 pp, medium confidence |
| Claude Opus 4.8 | 47.8% | 54.3% | estimated ± 6.9 pp, medium confidence |
| Claude Opus 5 | 61.2% | 62.7% | estimated ± 6.9 pp, medium confidence |
| Claude Sonnet 5 | 48.3% | 54.7% | estimated ± 6.9 pp, medium confidence |
| Command A+ | 0.9% | 1.7% | estimated ± 6.9 pp, low confidence |
| DeepSeek V3 | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| DeepSeek V3 0324 | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| DeepSeek V4 Flash 0731 | 46.9% | 53.7% | estimated ± 6.9 pp, medium confidence |
| DeepSeek V4 Pro 0813 | 54.5% | 58.8% | estimated ± 6.9 pp, medium confidence |
| Gemini 2.5 Pro | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Gemini 3.1 Pro | 14.7% | 22.9% | estimated ± 6.9 pp, low confidence |
| Gemini 3.5 Flash | 42.2% | 50.3% | estimated ± 6.9 pp, medium confidence |
| Gemini 3.5 Flash-Lite | 24.3% | 34.2% | estimated ± 6.9 pp, low confidence |
| Gemini 3.6 Flash | 39.3% | 48.0% | estimated ± 6.9 pp, medium confidence |
| Gemini 3.7 Flash | 44.6% | 52.1% | estimated ± 6.9 pp, medium confidence |
| Gemma 3 27B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Gemma 4 12B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Gemma 4 26B A4B | 3.4% | 6.1% | estimated ± 6.9 pp, low confidence |
| Gemma 4 31B | 6.1% | 10.5% | estimated ± 6.9 pp, low confidence |
| Gemma 4 E2B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Gemma 4 E4B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| GLM-4.7 | 25.8% | 35.7% | estimated ± 6.9 pp, low confidence |
| GLM-5.1 | 31.0% | 40.8% | estimated ± 6.9 pp, medium confidence |
| GLM-5.2 | 43.7% | 51.4% | estimated ± 6.9 pp, medium confidence |
| GPT-4.1 mini | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| GPT-4.1 nano | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| GPT-4o | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| GPT-4o mini | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| GPT-5.1 | 16.5% | 25.2% | estimated ± 6.9 pp, low confidence |
| GPT-5.4 | 37.4% | 46.5% | estimated ± 6.9 pp, medium confidence |
| GPT-5.4 mini | 25.8% | 35.7% | estimated ± 6.9 pp, low confidence |
| GPT-5.4 nano | 22.5% | 32.2% | estimated ± 6.9 pp, low confidence |
| GPT-5.5 | 42.7% | 50.7% | estimated ± 6.9 pp, medium confidence |
| GPT-5.6 Luna | 48.2% | 54.6% | estimated ± 6.9 pp, medium confidence |
| GPT-5.6 Sol | 55.6% | 59.4% | estimated ± 6.9 pp, medium confidence |
| GPT-5.6 Terra | 47.7% | 54.3% | estimated ± 6.9 pp, medium confidence |
| GPT-5 (high) | 21.0% | 30.6% | estimated ± 6.9 pp, low confidence |
| GPT-OSS 120B | 5.6% | 9.7% | estimated ± 6.9 pp, low confidence |
| GPT-OSS 20B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Granite 4.2 30B | 3.9% | 6.9% | estimated ± 6.9 pp, low confidence |
| Granite 4.2 3B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Granite 4.2 8B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Grok 4.3 | 29.2% | 39.1% | estimated ± 6.9 pp, medium confidence |
| Grok 4.5 | 44.5% | 52.0% | estimated ± 6.9 pp, medium confidence |
| Hy3 | 28.3% | 38.3% | estimated ± 6.9 pp, low confidence |
| Hy3 Preview | 35.8% | 45.1% | estimated ± 6.9 pp, medium confidence |
| Inkling-Small | 31.2% | 41.0% | estimated ± 6.9 pp, medium confidence |
| K-Exaone | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Kimi K2.6 | 27.0% | 37.0% | estimated ± 6.9 pp, low confidence |
| Kimi K2.5 | 17.2% | 26.1% | estimated ± 6.9 pp, low confidence |
| Kimi K2.5 (Reasoning) | 17.2% | 26.1% | estimated ± 6.9 pp, low confidence |
| Kimi K2.7 Code | 27.0% | 37.0% | estimated ± 6.9 pp, low confidence |
| LFM2.5-2.6B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Ling 2.6 Flash | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Ling 3.0 Flash | 22.4% | 32.1% | estimated ± 6.9 pp, low confidence |
| Ling 3.0 Flash FP8 | 22.4% | 32.1% | estimated ± 6.9 pp, low confidence |
| Ling 3.0 Flash VL | 33.2% | 42.9% | estimated ± 6.9 pp, medium confidence |
| Ling 3.0 Tiny | 3.2% | 5.7% | estimated ± 6.9 pp, low confidence |
| Ling 3.1 Flash | 56.1% | 59.7% | estimated ± 6.9 pp, medium confidence |
| Llama 4 Maverick | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Llama 4 Scout | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Mercury 2.5 | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| MiMo-V2.5-Pro | 31.2% | 41.0% | estimated ± 6.9 pp, medium confidence |
| MiMo-V2.6-Flash | 55.5% | 59.4% | estimated ± 6.9 pp, medium confidence |
| MiMo-V2-Flash | 6.6% | 11.3% | estimated ± 6.9 pp, low confidence |
| MiniCPM5-2B | 10.5% | 17.2% | estimated ± 6.9 pp, low confidence |
| MiniMax M2.7 | 25.7% | 35.6% | estimated ± 6.9 pp, low confidence |
| Mistral Large 3 | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Mistral Medium 3.5 128B | 13.2% | 20.9% | estimated ± 6.9 pp, low confidence |
| Mistral Small 4 | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Mistral Small 4 (Reasoning) | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Muse Spark | 25.1% | 35.0% | estimated ± 6.9 pp, low confidence |
| Muse Spark 1.1 | 35.7% | 45.0% | estimated ± 6.9 pp, medium confidence |
| Muse Spark 1.2 | 48.9% | 55.1% | estimated ± 6.9 pp, medium confidence |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 7.0% | 11.9% | estimated ± 6.9 pp, low confidence |
| Nemotron 3 Nano 30B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Nemotron 3 Nano Omni 30B A3B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Nemotron 3 Super 100B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| North Mini Code | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Quasar 438B | 34.2% | 43.7% | estimated ± 6.9 pp, medium confidence |
| Qwen3.5-122B-A10B | 15.8% | 24.3% | estimated ± 6.9 pp, low confidence |
| Qwen3.6-27B | 24.4% | 34.3% | estimated ± 6.9 pp, low confidence |
| Qwen3.6-35B-A3B | 19.8% | 29.2% | estimated ± 6.9 pp, low confidence |
| Qwen3.6 Plus | 24.7% | 34.6% | estimated ± 6.9 pp, low confidence |
| Qwen3.7 Max | 31.6% | 41.4% | estimated ± 6.9 pp, medium confidence |
| Qwen3.7 Plus | 13.5% | 21.3% | estimated ± 6.9 pp, low confidence |
| Qwen3.8-Flash-Next | 56.6% | 60.0% | estimated ± 6.9 pp, medium confidence |
| Qwen3.8 Max Preview | 58.6% | 61.2% | estimated ± 6.9 pp, medium confidence |
| Solar Pro 3 | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Solar Pro 4 | 30.5% | 40.4% | estimated ± 6.9 pp, medium confidence |
| Step 3.7 Flash | 25.8% | 35.7% | estimated ± 6.9 pp, low confidence |
| Trinity-Large-Preview | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Trinity-Large-Thinking | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |
| Ultravox v0.6 Llama 3.3 70B | 0.0% | 0.0% | estimated ± 6.9 pp, low confidence |