Video-MME-Logical: leaderboard
Video-MME-Logical (2026) tests temporal-logical reasoning in video with procedurally generated clips whose answers come from the generating program: state tracking, sequential counting, temporal ordering, dynamic spatial reasoning and structural composition, 3,750 test videos, exact-match scoring. The board is overall accuracy in percent; human annotators reach about 96, and higher is better.
Metric: Overall accuracy (%) over all test videos; Video-MME-Logical: 3,750 procedurally generated test videos in 25 task categories under five temporal-logical operations (state tracking, sequential counting, temporal ordering, dynamic spatiality, structural composition), each at easy, medium and hard settings; zero-shot, 2 fps; exact-match accuracy of the tagged multiple-choice or fill-in answer; human 95.9; higher is better. Source: arxiv.org. Saturation forecast: Around August 2027. 17 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 28.6 | #94 |
| 2 | GPT-5.4 | 22.7 | #123 |
| 3 | Qwen 2.5 VL 72B Instruct | 12.5 | #846 |
| 4 | Qwen 3 VL 8B Instruct | 11.9 | #981 |
| 5 | Qwen 3 VL 30B A3B Instruct | 11.8 | #871 |
| 6 | Qwen 3 VL 30B A3B (Thinking) | 10.3 | #712 |
| 7 | Qwen 2.5 VL 7B Instruct | 7.4 | #1833 |
| 8 | Qwen 3 VL 8B (Thinking) | 6.6 | #887 |
| 9 | Qwen3 Omni 30B A3B Instruct | 5.8 | #904 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=video-mme-logical · How It Works · Data refreshed daily, snapshot 2026-11-13.