Video-MME
We show this table for reference; we do not rank on it.
A comprehensive benchmark for multimodal large language models on video understanding, covering temporal reasoning, perception, and question answering over videos.
Benchmark score on Video-MME — November 13, 2026
We compile the Video-MME rows from provider self-reports and secondary reports. Seed 2.1 Pro leads the table at 89.2%, followed by Seed 2.1 Turbo (89.0%) and Kimi K2.5 (87.4%). We do not use these results to rank models overall.
Seed 2.1 Pro
ByteDance
Reported by ByteDanceSeed 2.1 Turbo
ByteDance
Reported by ByteDanceKimi K2.5
Moonshot AI
Carried from an upstream page4 modelsMultimodal & GroundedRefreshingDisplay onlyUpdated November 13, 2026
Benchmark score table (4 models)
None of the 4 results on this page has a documented run by the benchmark or an independent evaluator: 3 reported by the labs that built the models and 1 carried from an upstream page, where who ran it is not recorded.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Seed 2.1 ProByteDanceReported by ByteDance | 89.2% | Reported by ByteDance | Not reported | Closed |
| 2 | Seed 2.1 TurboByteDanceReported by ByteDance | 89.0% | Reported by ByteDance | Not reported | Closed |
| 3 | Kimi K2.5Moonshot AICarried from an upstream page | 87.4% | Carried from an upstream page | Not reported | Open |
| 4 | LongCat-2.5MeituanReported by Meituan | 85.2% | Reported by Meituan | Not reported | Open |
- Lab-reported
- Benchmark-run
- Independent
Among the reported Video-MME rows, Seed 2.1 Pro is first at 89.2%. The third row is 1.8 points behind. The broader top-10 range is 4.0 points, so many of the published results sit in a relatively narrow band.
4 models have been evaluated on Video-MME. The benchmark falls in the Multimodal & Grounded category. Video-MME is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Video-MME
Year
2024
Tasks
Video understanding
Format
Video QA and analysis
Difficulty
Broad multimodal video reasoning
BenchLM tracks the aggregate Video-MME row as a display-oriented video benchmark when providers publish a single overall score rather than separate with-subtitle and without-subtitle splits.
Freshness and provenance
Version
Video-MME 2024
Refresh cadence
Annual
Staleness state
Refreshing
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Video-MME measure?
A comprehensive benchmark for multimodal large language models on video understanding, covering temporal reasoning, perception, and question answering over videos.
Which model scores highest on Video-MME?
Seed 2.1 Pro by ByteDance currently leads with a score of 89.2% on Video-MME.
How many models are evaluated on Video-MME?
4 AI models have published results on Video-MME in the BenchLM catalog.
Compare top models on Video-MME
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.