Video-MME with subtitle (Video-MME (with subtitle))
We show this table for reference; we do not rank on it.
A video understanding benchmark that allows subtitle access when answering multimodal questions about videos.
Benchmark score on Video-MME (with subtitle) — November 13, 2026
We compile the Video-MME (with subtitle) rows from provider self-reports and secondary reports. Qwen3.8 Max leads the table at 90.4%, followed by Gemini 3 Pro (88.4%) and Qwen3.7 Plus (88.0%). We do not use these results to rank models overall.
Qwen3.8 Max
Alibaba
Reported by AlibabaGemini 3 Pro
Qwen3.7 Plus
Alibaba
Reported by Alibaba12 modelsMultimodal & GroundedCurrentDisplay onlyUpdated November 13, 2026
Benchmark score table (12 models)
None of the 12 results on this page has a documented run by the benchmark or an independent evaluator: 8 reported by the labs that built the models and 4 reported in another publisher’s document.
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Qwen3.8 MaxAlibabaReported by Alibaba | 90.4% | Reported by Alibaba | Not reported | Open |
| 2 | Gemini 3 ProGoogleAnother lab’s comparison table | 88.4% | Another lab’s comparison table | Not reported | Closed |
| 3 | Qwen3.7 PlusAlibabaReported by Alibaba | 88.0% | Reported by Alibaba | Not reported | Closed |
| 4 | Qwen3.6-27BAlibabaReported by Alibaba | 87.7% | Reported by Alibaba | Not reported | Open |
| 5 | MiMo-V2.5XiaomiReported by Xiaomi | 87.7% | Reported by Xiaomi | Not reported | Closed |
| 6 | Qwen3.5 397BAlibabaReported by Alibaba | 87.5% | Reported by Alibaba | Not reported | Open |
| 7 | Kimi K2.5Moonshot AIAnother lab’s comparison table | 87.4% | Another lab’s comparison table | Not reported | Open |
| 8 | Qwen3.6-35B-A3BAlibabaReported by Alibaba | 86.6% | Reported by Alibaba | Not reported | Open |
| 9 | GPT-5.2OpenAIAnother lab’s comparison table | 86.0% | Another lab’s comparison table | Not reported | Closed |
| 10 | MiniMax M3MiniMaxReported by MiniMax | 85.4% | Reported by MiniMax | Not reported | Open |
| 11 | Claude Opus 4.5AnthropicAnother lab’s comparison table | 77.6% | Another lab’s comparison table | Not reported | Closed |
| 12 | LFM3-VL-2BLiquidAIReported by LiquidAI | 69.0% | Reported by LiquidAI | Not reported | Open |
- Lab-reported
- Benchmark-run
- Independent
Among the reported Video-MME (with subtitle) rows, Qwen3.8 Max is first at 90.4%. The third row is 2.4 points behind. The broader top-10 range is 5.0 points, so many of the published results sit in a relatively narrow band.
12 models have been evaluated on Video-MME (with subtitle). The benchmark falls in the Multimodal & Grounded category. Video-MME (with subtitle) is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Video-MME (with subtitle)
Year
2026
Tasks
Video understanding
Format
Video QA with subtitle context
Difficulty
Multimodal video reasoning
The subtitle-enabled Video-MME setting measures how well a model combines video perception with textual cues from subtitles rather than relying on frames alone.
Freshness and provenance
Version
Video-MME (with subtitle) 2026
Refresh cadence
Quarterly
Staleness state
Current
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 (with subtitle) measure?
A video understanding benchmark that allows subtitle access when answering multimodal questions about videos.
Which model scores highest on Video-MME (with subtitle)?
Qwen3.8 Max by Alibaba currently leads with a score of 90.4% on Video-MME (with subtitle).
How many models are evaluated on Video-MME (with subtitle)?
12 AI models have published results on Video-MME (with subtitle) in the BenchLM catalog.
Compare top models on Video-MME (with subtitle)
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.