multimodal benchmark
Video-MME Leaderboard
Video-MME is the first-ever comprehensive evaluation benchmark of Multi-modal Large Language Models (MLLMs) in video analysis. It features 900 videos totaling 254 hours with 2,700 human-annotated question-answer pairs across 6 primary visual domains (Knowledge, Film & Television, Sports Competition, Life Record, Multilingual, and others) and 30 subfields. The benchmark evaluates models across diverse temporal dimensions (11 seconds to 1 hour), integrates multi-modal inputs including video frames, subtitles, and audio, and uses rigorous manual labeling by expert annotators for precise assessment.
Updated Aug 17, 2026
Models17
Model coverage17
MetricScore
EvidenceB
Video-MME Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelBY | Score89.2% | Percentile100.0% | Participants17 | EvidenceC | Evaluated |
| Rank02 | ModelBY | Score89.0% | Percentile93.8% | Participants17 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score88.0% | Percentile87.5% | Participants17 | EvidenceC | Evaluated |
| Rank04 | ModelXI | Score87.7% | Percentile81.3% | Participants17 | EvidenceC | Evaluated |
| Rank05 | ModelMA | Score87.4% | Percentile75.0% | Participants17 | EvidenceC | Evaluated |
| Rank06 | ModelMI | Score85.4% | Percentile68.8% | Participants17 | EvidenceC | Evaluated |
| Rank07 | ModelGO | Score84.8% | Percentile62.5% | Participants17 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score84.2% | Percentile56.3% | Participants17 | EvidenceC | Evaluated |
| Rank09 | ModelGO | Score78.6% | Percentile50.0% | Participants17 | EvidenceC | Evaluated |
| Rank10 | ModelAM Nova 2 OmniAmazon | Score77.9% | Percentile43.8% | Participants17 | EvidenceC | Evaluated |
| Rank11 | ModelGO | Score76.1% | Percentile37.5% | Participants17 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score74.5% | Percentile31.3% | Participants17 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score73.3% | Percentile25.0% | Participants17 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score71.8% | Percentile18.8% | Participants17 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score71.4% | Percentile12.5% | Participants17 | EvidenceC | Evaluated |
| Rank16 | ModelGO | Score66.2% | Percentile6.3% | Participants17 | EvidenceC | Evaluated |
| Rank17 | ModelMI | Score55.0% | Percentile0.0% | Participants17 | EvidenceC | Evaluated |
Video-MME Highlights
The leading models and scores on this benchmark.
Video-MME Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for Video-MME
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis video-mme AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01BYSeed 2.1 ProByteDance
- Score
- 89.2%
Strengths
- Ranks #1 of 17 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures Video-MME, not total model capability
- 02BYSeed 2.1 TurboByteDance
- Score
- 89.0%
Strengths
- Ranks #2 of 17 compared models
- 94th percentile on this benchmark
- C evidence result
Considerations
- This result measures Video-MME, not total model capability
- 03ACQwen3.7-PlusAlibaba Cloud / Qwen Team
- Score
- 88.0%
- Price
- $0.50 input / $3.0 output per 1M tokens
Strengths
- Ranks #3 of 17 compared models
- 88th percentile on this benchmark
- C evidence result
Considerations
- This result measures Video-MME, not total model capability
- 04XIMiMo-V2.5Xiaomi
- Score
- 87.7%
- Price
- $0.14 input / $0.28 output per 1M tokens
- Speed
- Up to 87 tok/s via Novita
Strengths
- Ranks #4 of 17 compared models
- 81th percentile on this benchmark
- C evidence result
Considerations
- This result measures Video-MME, not total model capability
- 05MAKimi K2.5Moonshot AI
- Score
- 87.4%
- Price
- $0.60 input / $3.0 output per 1M tokens
Strengths
- Ranks #5 of 17 compared models
- 75th percentile on this benchmark
- C evidence result
Considerations
- This result measures Video-MME, not total model capability
Selection summary
Best AI Models for Video-MME
Seed 2.1 Pro currently leads Video-MME with 89.2%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is Video-MME?
What Video-MME measures and how its scores work.
Video-MME is the first-ever comprehensive evaluation benchmark of Multi-modal Large Language Models (MLLMs) in video analysis. It features 900 videos totaling 254 hours with 2,700 human-annotated question-answer pairs across 6 primary visual domains (Knowledge, Film & Television, Sports Competition, Life Record, Multilingual, and others) and 30 subfields. The benchmark evaluates models across diverse temporal dimensions (11 seconds to 1 hour), integrates multi-modal inputs including video frames, subtitles, and audio, and uses rigorous manual labeling by expert annotators for precise assessment.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- Video-MME
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- video-mme|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about Video-MME.
Which model scores highest on Video-MME?
Seed 2.1 Pro is currently ranked first with 89.2%.
What does Video-MME measure?
Video-MME is the first-ever comprehensive evaluation benchmark of Multi-modal Large Language Models (MLLMs) in video analysis. It features 900 videos totaling 254 hours with 2,700 human-annotated question-answer pairs across 6 primary visual domains (Knowledge, Film & Television, Sports Competition, Life Record, Multilingual, and others) and 30 subfields. The benchmark evaluates models across diverse temporal dimensions (11 seconds to 1 hour), integrates multi-modal inputs including video frames, subtitles, and audio, and uses rigorous manual labeling by expert annotators for precise assessment.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
17 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.