List your company
ResearchFormerly WizWand

Long Video Understanding on VideoMME Overall (w/o sub.) 2025

75AccuracyGemini-1.5-Pro
38.0847.66557.2566.835Aug 28, 2025
Updated 5mo ago

Evaluation results

MethodLinks
Gemini-1.5-ProSize=-, Frames=0.5 fps2025.0875
GPT-4oSize=-, Frames=0.5 fps / 3842025.0871.9
Qwen2.5-VL-7BSize=7B, Frames=7682025.0865.1
Video-MTRSize=7B, Frames=802025.0862.7
Video-MTRSize=7B, Frames=642025.0862.2
Video-R1Size=7B, Frames=642025.0861.4
Qwen2.5-VL-7BSize=7B, Frames=802025.0859.5
Video-R1Size=7B, Frames=322025.0859.3
Video-MTRSize=7B, Frames=322025.0859
Qwen2.5-VL-7BSize=7B, Frames=642025.0858.4
LLaVA-OneVisionSize=7B, Frames=322025.0858.2
VideoMemAgentSize=-, Frames=722025.0857.4
VideoAgentSize=-, Frames=872025.0856
Video-XLSize=7B, Frames=128 / 2562025.0855.5
Qwen2.5-VL-7BSize=7B, Frames=322025.0853.6
LongVASize=7B, Frames=128 / 2562025.0852.6
DrVideo(GPT-4)Size=-, Frames=0.2/0.5 fps2025.0851.7
Video-CCAMSize=9B, Frames=962025.0850.3
VideoLLaMA2Size=8×7B, Frames=82025.0847.9
Video-LLaVASize=7B, Frames=82025.0839.9
VideoChat2Size=7B, Frames=162025.0839.5