A benchmark for evaluating multimodal models on video understanding tasks across multiple disciplines, emphasizing temporal reasoning and comprehension over video content.
BenchLM mirrors the published score view for MMVU. Kimi K2.5 leads the public snapshot at 80.4% , followed by Qwen3.5-122B-A10B (74.7%) and Qwen3.5-27B (73.3%). BenchLM does not use these results to rank models overall.
Kimi K2.5
Moonshot AI
Qwen3.5-122B-A10B
Alibaba
Qwen3.5-27B
Alibaba
The published MMVU snapshot is tightly clustered at the top: Kimi K2.5 sits at 80.4%, while the third row is only 7.1 points behind. The broader top-10 spread is 8.1 points, so many of the published scores sit in a relatively narrow band.
4 models have been evaluated on MMVU. The benchmark falls in the Multimodal & Grounded category. This category carries a 12% weight in BenchLM.ai's overall scoring system. MMVU is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
Year
2026
Tasks
Video understanding
Format
Video reasoning benchmark
Difficulty
Multi-disciplinary multimodal video reasoning
MMVU is a useful video-understanding benchmark for BenchLM because it appears in frontier provider tables and complements Video-MME with a more discipline-oriented video reasoning slice.
Version
MMVU 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.
A benchmark for evaluating multimodal models on video understanding tasks across multiple disciplines, emphasizing temporal reasoning and comprehension over video content.
Kimi K2.5 by Moonshot AI currently leads with a score of 80.4% on MMVU.
4 AI models have been evaluated on MMVU on BenchLM.
Get notified when new models drop, benchmark scores change, or the leaderboard shifts. One email per week.
Free. No spam. Unsubscribe anytime. We only store derived location metadata for consent routing.