Skip to content

Video-MME with subtitle

A video understanding benchmark that allows subtitle access when answering multimodal questions about videos.

11 models

Higher is better.

  1. 1Qwen3.8 MaxAlibaba90.4
  2. 2Gemini 3 ProGoogle88.4
  3. 3Qwen3.7 PlusAlibaba88.0
  4. 4Qwen3.6-27BAlibaba87.7
  5. 5MiMo-V2.5Xiaomi87.7
  6. 6Qwen3.5 397BAlibaba87.5
  7. 7Kimi K2.5Moonshot AI87.4
  8. 8Qwen3.6-35B-A3BAlibaba86.6
  9. 9GPT-5.2OpenAI86.0
  10. 10MiniMax M3MiniMax85.4
  11. 11Claude Opus 4.5Anthropic77.6
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Video QA with subtitle context
Tasks
Video understanding
Authors
Qwen (2026)

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.