Calibration
EnterpriseOps-Gym → AA-AnalystAgent
AA-AnalystAgent is estimated from EnterpriseOps-Gym with a inverse Michaelis–Menten curve fitted on 7 models measured on both: y = 3.31848·(x − 0.2538) / (0.2538 + 2.0000 − x), R² = 0.97, cross-validated error 3.6 pp. It is used for 8 estimates.
| Estimated model | EnterpriseOps-Gym | AA-AnalystAgent | Source |
|---|---|---|---|
| GLM-5.3-Flash | 33.2% | 13.5% | estimated ± 3.6 pp, medium confidence |
| GLM-5.3 (max) | 36.4% | 19.4% | estimated ± 3.6 pp, medium confidence |
| Qwen3.8 27B (xhigh) | 44.2% | 34.5% | estimated ± 3.6 pp, medium confidence |
| Muse Glimmer (high) | 34.7% | 16.2% | estimated ± 3.6 pp, medium confidence |
| Grok 4.6 (high) | 48.3% | 43.0% | estimated ± 3.6 pp, medium confidence |
| Gemini 3.7 Flash (medium) | 50.4% | 47.5% | estimated ± 3.6 pp, medium confidence |
| Gemini 3.5 Flash (high) | 50.1% | 46.8% | estimated ± 3.6 pp, medium confidence |
| DeepSeek V4 Pro 0813 (max) | 49.6% | 45.7% | estimated ± 3.6 pp, medium confidence |