Wikibench

MLVU

multimodal official site →

Multilingual: no

A comprehensive benchmark for multi-task long video understanding that evaluates multimodal large language models on videos ranging from 3 minutes to 2 hours across 9 distinct tasks including reasoning, captioning, recognition, and summarization.

Methodology

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

Leaderboard

Showing 8 of 8 results

  1. Qwen3.5-122B-A10B self-reported llm-stats
    87.3%
  2. Qwen3.6 Plus self-reported llm-stats
    86.7%
  3. Qwen3.6-27B self-reported llm-stats
    86.6%
  4. Qwen3.5-27B self-reported llm-stats
    85.9%
  5. Qwen3.5-35B-A3B self-reported llm-stats
    85.6%
  6. Qwen3 VL 235B A22B Instruct self-reported llm-stats
    84.3%
  7. Qwen3 VL 235B A22B Thinking self-reported llm-stats
    83.8%
  8. Qwen2.5 VL 7B Instruct self-reported llm-stats
    70.2%