Cad eval — Epoch AI
Cad eval — Epoch AI is run by Epoch AI. It has ranked 15 models, of which the catalogue holds 13, scoring from 0.12 to 0.74.
measured by Epoch AI · the board itself
| Place | Model | Metric | Score |
|---|---|---|---|
| 1stof 15 | o3 | Overall pass (%) (medium) | 0.74 |
| 2ndof 15 | Gemini 2.5 Pro | Overall pass (%) | 0.64 |
| 3rdof 15 | o4-mini | Overall pass (%) (medium) | 0.62 |
| 4thof 15 | o1 | Overall pass (%) (medium) | 0.56 |
| 5thof 15 | Claude 3.7 Sonnet | Overall pass (%) | 0.54 |
| 6thof 15 | o3-mini | Overall pass (%) (medium) | 0.54 |
| 7thof 15 | Claude 3.5 Sonnet | Overall pass (%) | 0.48 |
| 8thof 15 | GPT-4.1 fine-tuned | Overall pass (%) | 0.42 |
| 10thof 15 | Claude Haiku 3.5 | Overall pass (%) | 0.32 |
| 11thof 15 | gemini-2.0-flash-001 | Overall pass (%) | 0.3 |
| 12thof 15 | GPT-4o (2024-08-06) | Overall pass (%) | 0.26 |
| 14thof 15 | GPT-4.1 mini | Overall pass (%) | 0.16 |
| 15thof 15 | Claude 3 Haiku | Overall pass (%) | 0.12 |
Read from the board on 2026-09-16