List your company
ResearchFormerly WizWand

Video Question Answering on MMVU (Accuracy)

58.5AccuracyVanilla
43.62847.48951.3555.211Mar 30, 2026
Updated 7mo ago

Evaluation results

MethodLinks
VanillaBackbone=Qwen3-VL-8B, Retention Ratio R=100%, Reasoning (CoT)=false, Frames=128 Frames2026.0358.5
VanillaBackbone=Qwen3-VL-8B, Retention Ratio R=100%, Reasoning (CoT)=false, Frames=32 Frames2026.0357.5
Random DropBackbone=Qwen3-VL-8B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0357.1
Random DropBackbone=Qwen3-VL-8B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0356.8
VideoAuto-R1Backbone=Qwen2.5-VL-7B, Retention Ratio R=100%, Reasoning (CoT)=true, Frames=128 Frames2026.0356.7
ResAdaptBackbone=Qwen3-VL-8B, Retention Ratio R=22.9%, Reasoning (CoT)=false, Frames=128 Frames2026.0356.3
ToMeBackbone=Qwen3-VL-8B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0356
ToMeBackbone=Qwen3-VL-8B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0355.9
VisionZipBackbone=Qwen3-VL-8B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0355.7
Random DropBackbone=Qwen3-VL-8B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0355.4
ResAdaptBackbone=Qwen3-VL-8B, Retention Ratio R=11.1%, Reasoning (CoT)=false, Frames=128 Frames2026.0355.4
ResAdaptBackbone=Qwen3-VL-8B, Retention Ratio R=23.8%, Reasoning (CoT)=false, Frames=32 Frames2026.0355.3
ToMeBackbone=Qwen3-VL-8B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0355.1
VideoAuto-R1Backbone=Qwen2.5-VL-7B, Retention Ratio R=100%, Reasoning (CoT)=true, Frames=32 Frames2026.0355
FlashVidBackbone=Qwen3-VL-8B, Retention Ratio R=30.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0354.8
FlashVidBackbone=Qwen3-VL-8B, Retention Ratio R=12.2%, Reasoning (CoT)=false, Frames=32 Frames2026.0354.8
ResAdaptBackbone=Qwen3-VL-8B, Retention Ratio R=11.4%, Reasoning (CoT)=false, Frames=32 Frames2026.0354.6
VisionZipBackbone=Qwen3-VL-8B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0354.4
FixedScaleBackbone=Qwen3-VL-8B, Retention Ratio R=12.3%, Reasoning (CoT)=false, Frames=128 Frames2026.0354.4
VisionZipBackbone=Qwen3-VL-8B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0354.2
FixedScaleBackbone=Qwen3-VL-8B, Retention Ratio R=12.3%, Reasoning (CoT)=false, Frames=32 Frames2026.0353.8
VisionZipBackbone=Qwen3-VL-8B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0353.7
VideoAuto-R1 + ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=23.8%, Reasoning (CoT)=true, Frames=128 Frames2026.0353.5
VideoAuto-R1 + ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=23.8%, Reasoning (CoT)=true, Frames=32 Frames2026.0353.2
Random DropBackbone=Qwen3-VL-8B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0353.2
VanillaBackbone=Qwen2.5-VL-7B, Retention Ratio R=100%, Reasoning (CoT)=false, Frames=128 Frames2026.0353.1
ToMeBackbone=Qwen3-VL-8B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0353.1
ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=22.9%, Reasoning (CoT)=false, Frames=128 Frames2026.0352.8
VanillaBackbone=Qwen2.5-VL-7B, Retention Ratio R=100%, Reasoning (CoT)=false, Frames=32 Frames2026.0352.7
VideoAuto-R1 + ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=11.4%, Reasoning (CoT)=true, Frames=128 Frames2026.0352.4
ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=23.8%, Reasoning (CoT)=false, Frames=32 Frames2026.0351.9
ToMeBackbone=Qwen2.5-VL-7B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0351.9
VideoAuto-R1 + ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=11.4%, Reasoning (CoT)=true, Frames=32 Frames2026.0351.8
FixedScaleBackbone=Qwen2.5-VL-7B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0351.2
FlashVidBackbone=Qwen2.5-VL-7B, Retention Ratio R=29.3%, Reasoning (CoT)=false, Frames=32 Frames2026.0351.1
VisionZipBackbone=Qwen2.5-VL-7B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0351.1
ToMeBackbone=Qwen2.5-VL-7B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0351
Random DropBackbone=Qwen2.5-VL-7B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0350.8
VisionZipBackbone=Qwen2.5-VL-7B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0349.8
Random DropBackbone=Qwen2.5-VL-7B, Retention Ratio R=25.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0349.6
FixedScaleBackbone=Qwen2.5-VL-7B, Retention Ratio R=12.3%, Reasoning (CoT)=false, Frames=128 Frames2026.0349.6
ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=11.4%, Reasoning (CoT)=false, Frames=32 Frames2026.0349.2
ResAdaptBackbone=Qwen2.5-VL-7B, Retention Ratio R=11.1%, Reasoning (CoT)=false, Frames=128 Frames2026.0349
ToMeBackbone=Qwen2.5-VL-7B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0348.9
FlashVidBackbone=Qwen2.5-VL-7B, Retention Ratio R=10.4%, Reasoning (CoT)=false, Frames=32 Frames2026.0347.9
FixedScaleBackbone=Qwen2.5-VL-7B, Retention Ratio R=12.3%, Reasoning (CoT)=false, Frames=32 Frames2026.0347.7
VisionZipBackbone=Qwen2.5-VL-7B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0347.6
Random DropBackbone=Qwen2.5-VL-7B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=32 Frames2026.0347.1
Random DropBackbone=Qwen2.5-VL-7B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0345.8
VisionZipBackbone=Qwen2.5-VL-7B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0344.8
ToMeBackbone=Qwen2.5-VL-7B, Retention Ratio R=10.0%, Reasoning (CoT)=false, Frames=128 Frames2026.0344.2