Motion Understanding and Temporal Reasoning on Tomato (test)
53AccuracyGPT-5
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5type=Proprietary2026.02 | 53 | |
| Gemini 2.5 Protype=Proprietary2026.02 | 48.6 | |
| Gemini 3 Protype=Proprietary2026.02 | 48.3 | |
| Claude Sonnet 4.5type=Proprietary2026.02 | 39.6 | |
| CoPE-VideoLMnumber of parameters=7B2026.02 | 28.3 | |
| VideoCCAM-9Bnumber of parameters=9B2026.02 | 27 | |
| LLaVA-OV-7Bnumber of parameters=7B2026.02 | 25.5 | |
| LLaVA-Video-7Bnumber of parameters=7B2026.02 | 24.9 | |
| InternVL2-8Bnumber of parameters=8B2026.02 | 21.7 | |
| VideoLLaMA2-7Bnumber of parameters=7B2026.02 | 18.5 |