VideoMMMU

A video extension of MMMU-style multimodal reasoning over expert questions grounded in temporal media.

Not in index
Multimodal
Best in multimodal

Models13

Top result—

Top-3 spread—

Year—

Result and price

10 of 13 results

Results

13 results

#ModelResult

—Qwen3.8 Max88.7%

—Gemini 3 Pro87.6%

—dots3-note Preview86.8%

—Kimi K2.586.6%

—GPT-5.285.9%

—Qwen3.7 Plus85.4%

—Qwen3.5 397B84.7%

—MiniMax M384.6%

—Qwen3.6-27B84.4%

—Claude Opus 4.584.4%

—Qwen3.6 Plus84.0%

—Qwen3.6-35B-A3B83.7%

—LFM3-VL-2B49.6%