Calibration
SkillsBench official leaderboard (BenchFlow) → EBR-bench
EBR-bench is estimated from SkillsBench official leaderboard (BenchFlow) with a linear curve fitted on 10 models measured on both: y = 0.8477·x + -0.2958, R² = 0.45, cross-validated error 9.1 pp. It is used for 10 estimates.
| Estimated model | SkillsBench official leaderboard (BenchFlow) | EBR-bench | Source |
|---|---|---|---|
| GLM-5.1 | 58.4% | 19.9% | estimated ± 9.1 pp, low confidence |
| Hy3 | 55.9% | 17.8% | estimated ± 9.1 pp, low confidence |
| Gemini 3 Flash | 54.6% | 16.7% | estimated ± 9.1 pp, low confidence |
| MiniMax M3 | 53.0% | 15.4% | estimated ± 9.1 pp, low confidence |
| Claude Sonnet 4.6 | 47.2% | 10.4% | estimated ± 9.1 pp, low confidence |
| Grok 4.3 | 41.7% | 5.8% | estimated ± 9.1 pp, low confidence |
| GPT-5.4 mini | 41.4% | 5.5% | estimated ± 9.1 pp, low confidence |
| MiniMax M2.7 | 34.9% | 0.0% | estimated ± 9.1 pp, low confidence |
| Claude Haiku 4.5 | 30.1% | 0.0% | estimated ± 9.1 pp, low confidence |
| Gemini 3.1 Flash-Lite | 20.1% | 0.0% | estimated ± 9.1 pp, low confidence |
Fitted on
| Model | SkillsBench official leaderboard (BenchFlow) | EBR-bench |
|---|---|---|
| Claude Opus 4.5 | 49.0% | 14.3% |
| Claude Opus 4.6 | 50.2% | 12.7% |
| Claude Opus 4.7 | 61.2% | 19.0% |
| Claude Opus 4.8 | 54.1% | 28.6% |
| Claude Sonnet 4.5 | 36.2% | 2.4% |
| Gemini 3.1 Pro | 60.8% | 14.3% |
| Gemini 3.5 Flash | 48.2% | 4.8% |
| GPT-5.2 | 51.7% | 23.0% |
| GPT-5.5 | 67.3% | 34.3% |
| Kimi K2.6 | 54.0% | 2.4% |