Wikibench

VideoMMMU

reasoning official site →

Multilingual: no

Video-MMMU evaluates Large Multimodal Models' ability to acquire knowledge from expert-level professional videos across six disciplines through three cognitive stages: perception, comprehension, and adaptation. Contains 300 videos and 900 human-annotated questions spanning Art, Business, Science, Medicine, Humanities, and Engineering.

Methodology

Imported from llm-stats public benchmark metadata. Modality: multimodal. Max score: 1. Categories: healthcare, multimodal, reasoning, vision. Language: en. Verified by llm-stats: no.

Leaderboard

Showing 20 of 24 results

  1. Gemini 3 Pro self-reported llm-stats
    87.6%
  2. Gemini 3 Flash self-reported llm-stats
    86.9%
  3. Kimi K2.5 self-reported llm-stats
    86.6%
  4. GPT-5.2 self-reported llm-stats
    85.9%
  5. Gemini 3.1 Flash-Lite self-reported llm-stats
    84.8%
  6. GPT-5 self-reported llm-stats
    84.6%
  7. MiniMax M3 self-reported llm-stats
    84.6%
  8. Qwen3.6-27B self-reported llm-stats
    84.4%
  9. Qwen3.6 Plus self-reported llm-stats
    84.0%
  10. Gemini 2.5 Pro Preview 06-05 self-reported llm-stats
    83.6%
  11. o3 self-reported llm-stats
    83.3%
  12. Qwen3.5-27B self-reported llm-stats
    82.3%
  13. Qwen3.5-122B-A10B self-reported llm-stats
    82.0%
  14. Qwen3.5-35B-A3B self-reported llm-stats
    80.4%
  15. Qwen3 VL 235B A22B Thinking self-reported llm-stats
    80.0%
  16. Qwen3 VL 32B Thinking self-reported llm-stats
    79.0%
  17. Qwen3 VL 30B A3B Thinking self-reported llm-stats
    75.0%
  18. Qwen3 VL 235B A22B Instruct self-reported llm-stats
    74.7%
  19. Qwen3 VL 8B Thinking self-reported llm-stats
    72.8%
  20. Qwen3 VL 4B Thinking self-reported llm-stats
    69.4%