List your company

Video Reasoning Quality Assessment on UrbanVideo and VSI-Bench Clean

3.91Original ScoreGemini-1.5-Pro
2.83883.11693.3953.6731Mar 11, 2026

Evaluation results

MethodLinks
Gemini-1.5-ProSize=–, #Fr=322026.033.91
GPT-4oSize=–, #Fr=322026.033.82
ROVASize=72B, #Fr=32, Backbone=Qwen2.5-VL2026.033.72
Video-R1Size=72B, #Fr=322026.033.68
Claude-3.5-Son.Size=–, #Fr=322026.033.65
Qwen2.5-VLSize=72B, #Fr=322026.033.64
ROVASize=13B, #Fr=32, Backbone=Qwen2-VL2026.033.62
ROVASize=7B, #Fr=32, Backbone=Embodied-R2026.033.58
ROVASize=7B, #Fr=32, Backbone=Qwen2.5-VL2026.033.52
Qwen2-VLSize=13B, #Fr=322026.033.48
Embodied-RSize=7B, #Fr=322026.033.45
Video-R1Size=7B, #Fr=322026.033.42
Qwen2.5-VLSize=7B, #Fr=322026.033.41
LLaVA-Video-RSize=7B, #Fr=322026.033.32
ROVASize=8B, #Fr=32, Backbone=InternVL2.52026.033.28
VideoChat-RSize=7B, #Fr=162026.033.25
MiniCPM-V 2.6Size=8B, #Fr=642026.033.18
InternVL2.5Size=8B, #Fr=322026.033.15
LLaVA-VideoSize=7B, #Fr=322026.033.12
VideoLLaMA2Size=7B, #Fr=162026.033.01
VideoChat2Size=7B, #Fr=162026.032.88