List your company
ResearchFormerly WizWand

Video Understanding on LongVideoBench 1-60min

56.8AccuracyCaCoVID
44.52847.71450.954.086Feb 2, 2026
Updated 8mo ago

Evaluation results

MethodLinks
CaCoVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=15%2026.0256.8
CaCoVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=20%2026.0256.5
LLaVA-OneVision-7BBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=100%2026.0256.4
DivPruneBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=25%2026.0256.4
CaCoVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=25%2026.0256.2
PruneVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=25%2026.0256.1
VisionZipBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=25%2026.0256
FrameFusionBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=25%2026.0255.7
VisionZipBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=20%2026.0255.7
DivPruneBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=20%2026.0255.6
CaCoVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=10%2026.0255.2
FrameFusionBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=20%2026.0255.1
DivPruneBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=15%2026.0255.1
PruneVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=15%2026.0255.1
PruneVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=20%2026.0254.9
PruneVIDBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=10%2026.0254.4
Qwen2.5-VL-3BBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=100%2026.0254.3
VisionZipBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=15%2026.0254
FastVBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=25%2026.0253.5
DivPruneBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=25%2026.0253.4
DivPruneBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=10%2026.0253.3
CaCoVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=25%2026.0253.3
FastVBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=20%2026.0253
PruneVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=25%2026.0253
FrameFusionBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=15%2026.0252.9
CaCoVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=20%2026.0252.8
FrameFusionBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=25%2026.0252.4
CaCoVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=15%2026.0252.4
DivPruneBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=20%2026.0252.3
PruneVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=20%2026.0252.1
PruneVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=15%2026.0252.1
VisionZipBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=15%2026.0252
VisionZipBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=20%2026.0251.9
VisionZipBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=25%2026.0251.7
DivPruneBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=15%2026.0251.5
VisionZipBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=10%2026.0251.5
CaCoVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=10%2026.0251.5
PruneVIDBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=10%2026.0251.3
FrameFusionBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=20%2026.0250.7
DivPruneBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=10%2026.0250.5
FrameFusionBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=15%2026.0249.9
FastVBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=15%2026.0249.8
FastVBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=25%2026.0249.5
FrameFusionBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=10%2026.0249.1
VisionZipBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=10%2026.0248.9
FastVBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=20%2026.0248.2
FastVBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=15%2026.0247.8
FastVBase Model=LLaVA-OneVision-7B, Equivalent Retention Ratio=10%2026.0247.6
FastVBase Model=Qwen2.5-VL-3B, Equivalent Retention Ratio=10%2026.0245