Calibration
Gert Labs → OSWorld-Verified
OSWorld-Verified is estimated from Gert Labs with a Hill curve fitted on 13 models measured on both: y = 0.5161 + (0.9170 − 0.5161)·x^6.00 / (0.61788^6.00 + x^6.00), R² = 0.79, cross-validated error 5.9 pp. It is used for 31 estimates.
| Estimated model | Gert Labs | OSWorld-Verified | Source |
|---|---|---|---|
| Claude 4 Sonnet | 39.7% | 54.2% | estimated ± 5.9 pp, medium confidence |
| DeepSeek V3.2 | 29.6% | 52.1% | estimated ± 5.9 pp, medium confidence |
| Gemini 2.5 Pro | 42.0% | 55.2% | estimated ± 5.9 pp, medium confidence |
| Gemini 3.1 Flash-Lite | 38.5% | 53.8% | estimated ± 5.9 pp, medium confidence |
| Gemini 3.1 Pro | 56.9% | 66.8% | estimated ± 5.9 pp, medium confidence |
| Gemini 3 Flash | 56.6% | 66.5% | estimated ± 5.9 pp, medium confidence |
| Gemini 3 Pro | 63.2% | 73.0% | estimated ± 5.9 pp, medium confidence |
| Gemma 4 31B | 35.3% | 53.0% | estimated ± 5.9 pp, medium confidence |
| GLM-5V-Turbo | 30.8% | 52.2% | estimated ± 5.9 pp, medium confidence |
| GPT-4.1 | 25.7% | 51.8% | estimated ± 5.9 pp, low confidence |
| GPT-5.1 | 41.2% | 54.9% | estimated ± 5.9 pp, medium confidence |
| GPT-5.1-Codex | 49.7% | 60.1% | estimated ± 5.9 pp, medium confidence |
| GPT-5.2-Codex | 51.8% | 61.9% | estimated ± 5.9 pp, medium confidence |
| GPT-OSS 120B | 29.6% | 52.1% | estimated ± 5.9 pp, medium confidence |
| Grok 4 | 42.3% | 55.4% | estimated ± 5.9 pp, medium confidence |
| Grok 4.1 Fast | 47.3% | 58.3% | estimated ± 5.9 pp, medium confidence |
| Grok 4.20 | 38.4% | 53.8% | estimated ± 5.9 pp, medium confidence |
| Grok 4.3 | 43.9% | 56.2% | estimated ± 5.9 pp, medium confidence |
| Grok Build 0.1 | 49.2% | 59.7% | estimated ± 5.9 pp, medium confidence |
| Hy3 Preview | 36.9% | 53.4% | estimated ± 5.9 pp, medium confidence |
| MiMo-V2.5 | 46.9% | 58.0% | estimated ± 5.9 pp, medium confidence |
| MiMo-V2.5-Pro | 62.7% | 72.5% | estimated ± 5.9 pp, medium confidence |
| MiMo-V2-Pro | 36.7% | 53.3% | estimated ± 5.9 pp, medium confidence |
| MiniMax M2.7 | 40.4% | 54.5% | estimated ± 5.9 pp, medium confidence |
| Mistral Medium 3.5 128B | 39.1% | 54.0% | estimated ± 5.9 pp, medium confidence |
| Qwen3.6-27B | 54.8% | 64.8% | estimated ± 5.9 pp, medium confidence |
| Qwen3.6-35B-A3B | 42.7% | 55.5% | estimated ± 5.9 pp, medium confidence |
| Qwen3.6 Plus | 50.6% | 60.9% | estimated ± 5.9 pp, medium confidence |
| Qwen3.7 Max | 64.3% | 74.0% | estimated ± 5.9 pp, medium confidence |
| Qwen3 Max | 43.7% | 56.1% | estimated ± 5.9 pp, medium confidence |
| Trinity-Large-Thinking | 32.6% | 52.5% | estimated ± 5.9 pp, medium confidence |