Official benchmarks map / Video-MME-v2
MiniMaxAI/MiniMax-M3 is ranked first on the Video-MME-v2 leaderboard with 85.4, 38.9% ahead of meta-models/Muse-Spark-Open-816B-A42B. Read live from the Hugging Face API on November 6, 2026.
| # | Model | Organization | Score |
|---|---|---|---|
| 1 | MiniMaxAI/MiniMax-M3 | MiniMaxAI | 85.4 |
| 2 | meta-models/Muse-Spark-Open-816B-A42B | meta-models | 61.5 |
| 3 | moonshotai/Kimi-K2.5 | moonshotai | 61.1 |
| 4 | Qwen/Qwen3.5-397B-A17B | Qwen | 55.9 |
| 5 | moonshotai/Kimi-K2.5 | moonshotai | 54.4 |
| 6 | Qwen/Qwen3.5-397B-A17B | Qwen | 48.9 |
| 7 | dots-studio/dots3-note-prev | dots-studio | 39.3 |
Video-MME-v2 is a dataset that Hugging Face marks as an official benchmark. Its leaderboard is built automatically from the .eval_results files that model repositories publish, so entries are mostly self-reported by model authors and evaluation settings differ between submissions. A verified tag appears only where Hugging Face marks an entry as verified.
Original board: huggingface.co/datasets/MME-Benchmarks/Video-MME-v2
Part of the Hugging Face official benchmarks map by QuantID. Unofficial. Not affiliated with or endorsed by Hugging Face.