Calibration
Gert Labs → ApprenticeBench
ApprenticeBench is estimated from Gert Labs with a linear curve fitted on 5 models measured on both: y = 1.4971·x + -0.8927, R² = 0.87, cross-validated error 3.1 pp. It is used for 47 estimates.
| Estimated model | Gert Labs | ApprenticeBench | Source |
|---|---|---|---|
| Claude 4 Sonnet | 39.7% | 0.0% | estimated ± 3.1 pp, low confidence |
| Claude Opus 4.5 | 64.2% | 6.9% | estimated ± 3.1 pp, medium confidence |
| Claude Opus 4.8 | 73.0% | 20.0% | estimated ± 3.1 pp, low confidence |
| Claude Sonnet 4.5 | 48.5% | 0.0% | estimated ± 3.1 pp, low confidence |
| DeepSeek V3.2 | 29.6% | 0.0% | estimated ± 3.1 pp, low confidence |
| Gemini 2.5 Pro | 42.0% | 0.0% | estimated ± 3.1 pp, low confidence |
| Gemini 3.1 Flash-Lite | 38.5% | 0.0% | estimated ± 3.1 pp, low confidence |
| Gemini 3.1 Pro | 56.9% | 0.0% | estimated ± 3.1 pp, low confidence |
| Gemini 3.5 Flash | 61.9% | 3.3% | estimated ± 3.1 pp, medium confidence |
| Gemini 3 Flash | 56.6% | 0.0% | estimated ± 3.1 pp, low confidence |
| Gemini 3 Pro | 63.2% | 5.4% | estimated ± 3.1 pp, medium confidence |
| Gemma 4 31B | 35.3% | 0.0% | estimated ± 3.1 pp, low confidence |
| GLM-4.7 | 40.0% | 0.0% | estimated ± 3.1 pp, low confidence |
| GLM-5 | 51.0% | 0.0% | estimated ± 3.1 pp, low confidence |
| GLM-5.1 | 60.1% | 0.7% | estimated ± 3.1 pp, low confidence |
| GLM-5V-Turbo | 30.8% | 0.0% | estimated ± 3.1 pp, low confidence |
| GPT-4.1 | 25.7% | 0.0% | estimated ± 3.1 pp, low confidence |
| GPT-5.1 | 41.2% | 0.0% | estimated ± 3.1 pp, low confidence |
| GPT-5.1-Codex | 49.7% | 0.0% | estimated ± 3.1 pp, low confidence |
| GPT-5.2 | 46.5% | 0.0% | estimated ± 3.1 pp, low confidence |
| GPT-5.2-Codex | 51.8% | 0.0% | estimated ± 3.1 pp, low confidence |
| GPT-5.3 Codex | 57.5% | 0.0% | estimated ± 3.1 pp, low confidence |
| GPT-OSS 120B | 29.6% | 0.0% | estimated ± 3.1 pp, low confidence |
| Grok 4 | 42.3% | 0.0% | estimated ± 3.1 pp, low confidence |
| Grok 4.1 Fast | 47.3% | 0.0% | estimated ± 3.1 pp, low confidence |
| Grok 4.20 | 38.4% | 0.0% | estimated ± 3.1 pp, low confidence |
| Grok 4.3 | 43.9% | 0.0% | estimated ± 3.1 pp, low confidence |
| Grok Build 0.1 | 49.2% | 0.0% | estimated ± 3.1 pp, low confidence |
| Hy3 Preview | 36.9% | 0.0% | estimated ± 3.1 pp, low confidence |
| Kimi K2.6 | 56.8% | 0.0% | estimated ± 3.1 pp, low confidence |
| Kimi K2.5 | 45.9% | 0.0% | estimated ± 3.1 pp, low confidence |
| Kimi K2.5 (Reasoning) | 32.6% | 0.0% | estimated ± 3.1 pp, low confidence |
| MiMo-V2.5 | 46.9% | 0.0% | estimated ± 3.1 pp, low confidence |
| MiMo-V2.5-Pro | 62.7% | 4.6% | estimated ± 3.1 pp, medium confidence |
| MiMo-V2-Pro | 36.7% | 0.0% | estimated ± 3.1 pp, low confidence |
| MiniMax M2.7 | 40.4% | 0.0% | estimated ± 3.1 pp, low confidence |
| Mistral Medium 3.5 128B | 39.1% | 0.0% | estimated ± 3.1 pp, low confidence |
| Qwen3.5-27B | 39.4% | 0.0% | estimated ± 3.1 pp, low confidence |
| Qwen3.5-35B-A3B | 29.0% | 0.0% | estimated ± 3.1 pp, low confidence |
| Qwen3.5 397B | 46.8% | 0.0% | estimated ± 3.1 pp, low confidence |
| Qwen3.6-27B | 54.8% | 0.0% | estimated ± 3.1 pp, low confidence |
| Qwen3.6-35B-A3B | 42.7% | 0.0% | estimated ± 3.1 pp, low confidence |
| Qwen3.6 Plus | 50.6% | 0.0% | estimated ± 3.1 pp, low confidence |
| Qwen3.7 Max | 64.3% | 7.0% | estimated ± 3.1 pp, medium confidence |
| Qwen3 Max | 43.7% | 0.0% | estimated ± 3.1 pp, low confidence |
| Step 3.7 Flash | 51.6% | 0.0% | estimated ± 3.1 pp, low confidence |
| Trinity-Large-Thinking | 32.6% | 0.0% | estimated ± 3.1 pp, low confidence |