Physical Perception on FysicsEval
40.3Prediction ScoreGPT-5
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GPT-5Size=-2026.02 | 40.3 | 3.48 | 89.9 | |
| Claude-4.5-HaikuSize=-2026.02 | 35.3 | 2.89 | 60.3 | |
| OmniFysicsSize=3B2026.02 | 32.6 | 3.22 | 94.7 | |
| SAIL-VL2Size=2B2026.02 | 21.9 | 2.58 | 84.7 | |
| InternVL3.5-8BSize=8B2026.02 | 21.7 | 2.53 | 80.7 | |
| Ovis2.5Size=2B2026.02 | 20.4 | 2.46 | 89.5 | |
| Qwen3-VL-8B-InstructSize=8B2026.02 | 20.1 | 2.65 | 90.1 | |
| Gemini-2.5-flashSize=-2026.02 | 19.8 | 3.1 | 89.4 | |
| Qwen2.5-OmniSize=3B2026.02 | 18.1 | 1.71 | 87.5 |