Calibration
Terminal-Bench 2.1 → Terminal-Bench 3.0
Terminal-Bench 3.0 is estimated from Terminal-Bench 2.1 with a inverse Michaelis–Menten curve fitted on 11 models measured on both: y = 0.03742·(x − 0.0000) / (0.0000 + 1.0234 − x), R² = 0.40, cross-validated error 9.1 pp. It is used for 33 estimates.
| Estimated model | Terminal-Bench 2.1 | Terminal-Bench 3.0 | Source |
|---|---|---|---|
| A.X K2 | 36.0% | 2.0% | estimated ± 9.1 pp, low confidence |
| Apodex 1.1 | 70.8% | 8.4% | estimated ± 9.1 pp, low confidence |
| Atria Dawn Preview | 78.3% | 12.2% | estimated ± 9.1 pp, low confidence |
| Claude Mythos 5 | 88.0% | 23.0% | estimated ± 9.1 pp, low confidence |
| dots3-note Preview | 75.1% | 10.3% | estimated ± 9.1 pp, low confidence |
| Ember-1 | 82.0% | 15.1% | estimated ± 9.1 pp, low confidence |
| Granite 4.2 30B | 29.2% | 1.5% | estimated ± 9.1 pp, low confidence |
| Granite 4.2 8B | 20.6% | 0.9% | estimated ± 9.1 pp, low confidence |
| K-EXAONE 2.0 | 43.8% | 2.8% | estimated ± 9.1 pp, low confidence |
| Laguna S 2.1 | 70.2% | 8.2% | estimated ± 9.1 pp, low confidence |
| MAI-Code-1.1-Flash | 62.9% | 6.0% | estimated ± 9.1 pp, low confidence |
| MiMo-V2.6-Flash | 87.6% | 22.2% | estimated ± 9.1 pp, low confidence |
| MiMo-V2.6-Pro | 89.9% | 27.0% | estimated ± 9.1 pp, low confidence |
| MiniCPM5-2B | 8.6% | 0.3% | estimated ± 9.1 pp, low confidence |
| Muse Glimmer 30B | 51.7% | 3.8% | estimated ± 9.1 pp, low confidence |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 23.5% | 1.1% | estimated ± 9.1 pp, low confidence |
| Ornith-1.0-35B | 64.2% | 6.3% | estimated ± 9.1 pp, low confidence |
| Ornith-1.0-397B | 77.5% | 11.7% | estimated ± 9.1 pp, low confidence |
| Ornith-1.0-9B | 43.1% | 2.7% | estimated ± 9.1 pp, low confidence |
| Ornith-1.5-9B | 46.2% | 3.1% | estimated ± 9.1 pp, low confidence |
| Pokee-Isaac 28B | 65.1% | 6.5% | estimated ± 9.1 pp, low confidence |
| Quasar 438B | 69.3% | 7.8% | estimated ± 9.1 pp, low confidence |
| Beam | 80.1% | 13.5% | estimated ± 9.1 pp, low confidence |
| Sakana Fugu | 80.2% | 13.6% | estimated ± 9.1 pp, low confidence |
| Sakana Fugu-Ultra | 82.1% | 15.2% | estimated ± 9.1 pp, low confidence |
| Seed 2.1 Pro | 71.0% | 8.5% | estimated ± 9.1 pp, low confidence |
| Seed 2.1 Turbo | 67.6% | 7.3% | estimated ± 9.1 pp, low confidence |
| Solar Pro 4 | 57.0% | 4.7% | estimated ± 9.1 pp, low confidence |
| Step 3.7 Flash | 59.5% | 5.2% | estimated ± 9.1 pp, low confidence |
| Step 5 Preview | 85.0% | 18.3% | estimated ± 9.1 pp, low confidence |
| SWE-1.7 | 81.5% | 14.6% | estimated ± 9.1 pp, low confidence |
| SWE-2 | 92.8% | 36.4% | estimated ± 9.1 pp, low confidence |
| Ternary Bonsai 2 27B | 52.8% | 4.0% | estimated ± 9.1 pp, low confidence |