Calibration
CyberGym → Gert Labs
Gert Labs is estimated from CyberGym with a linear curve fitted on 7 models measured on both: y = 0.3610·x + 0.3909, R² = 0.60, cross-validated error 6.0 pp. It is used for 16 estimates.
| Estimated model | CyberGym | Gert Labs | Source |
|---|---|---|---|
| Atria Dawn Preview | 86.5% | 70.3% | estimated ± 6.0 pp, low confidence |
| Claude Mythos 5 | 83.8% | 69.3% | estimated ± 6.0 pp, low confidence |
| Claude Mythos Preview | 83.1% | 69.1% | estimated ± 6.0 pp, low confidence |
| DeepSeek V4.1 Flash | 88.1% | 70.9% | estimated ± 6.0 pp, low confidence |
| DeepSeek V4 Flash 0731 | 76.7% | 66.8% | estimated ± 6.0 pp, low confidence |
| DeepSeek V4 Pro 0813 | 83.3% | 69.2% | estimated ± 6.0 pp, low confidence |
| Gemini 3.5 Flash Cyber | 83.2% | 69.1% | estimated ± 6.0 pp, low confidence |
| Gemini 3.8 Flash Cyber | 86.2% | 70.2% | estimated ± 6.0 pp, low confidence |
| GLM-5.3 | 84.5% | 69.6% | estimated ± 6.0 pp, low confidence |
| Hy4 preview | 78.4% | 67.4% | estimated ± 6.0 pp, low confidence |
| Ling 3.1 Flash | 87.9% | 70.8% | estimated ± 6.0 pp, low confidence |
| MiMo-V2.6-Flash | 95.1% | 73.4% | estimated ± 6.0 pp, low confidence |
| MiMo-V2.6-Pro | 94.0% | 73.0% | estimated ± 6.0 pp, low confidence |
| Muse Spark | 43.5% | 54.8% | estimated ± 6.0 pp, low confidence |
| Fugu Cyber | 86.9% | 70.5% | estimated ± 6.0 pp, low confidence |
| Step 5 Preview | 84.7% | 69.7% | estimated ± 6.0 pp, low confidence |