LongVideoBench

multimodal official site →

LongVideoBench is a question-answering benchmark featuring video-language interleaved inputs up to an hour long. It includes 3,763 varying-length web-collected videos with subtitles across diverse themes and 6,678 human-annotated multiple-choice questions in 17 fine-grained categories for comprehensive evaluation of long-term multimodal understanding.

Methodology

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

Leaderboard

  1. Kimi K2.5 self-reported llm-stats
    79.8%
  2. Qwen2.5 VL 7B Instruct self-reported llm-stats
    54.7%