List your company
ResearchFormerly WizWand

Video Question Answering on STAR

78.6AccuracyInternVideo2
15.097631.583848.0764.5562Sep 27, 2023Feb 17, 2024Jul 9, 2024Nov 30, 2024Apr 22, 2025Sep 12, 2025Feb 3, 2026
Updated 7mo ago

Evaluation results

MethodLinks
InternVideo2Params=6B2025.0378.6
VITED (Temporal Evidence Distillation)Params=7B, Base Model=TimeChat Base2025.0370.99
Open-o3-Video-7BParameters=7B2025.1070.5
Open-o3-Video-7BParameters=7B, Ablation=without Adaptive frame sampling2025.1070.1
VITED (Dense Caption Distillation)Params=7B, Base Model=TimeChat Base2025.0369.99
Open-o3-Video-7BParameters=7B, Ablation=without Gated mechanism2025.1069.6
TimeChat (Video Instruction Tuning)Params=7B, Base Model=TimeChat Base2025.0368.39
VITED (Temporal Evidence Distillation)Params=7B, Base Model=LLaVA-Video Base2025.0368.23
LLaVA-Video (Video Instruction Tuning)Params=7B, Base Model=LLaVA-Video Base2025.0367.74
VITED (Dense Caption Distillation)Params=7B, Base Model=LLaVA-Video Base2025.0367.56
Qwen2.5-VL-7BParameters=7B2025.1067.3
LLaVA-VideoParams=7B, Base Model=LLaVA-Video Base2025.0366.88
LLaVA-OneVisionParams=7B2025.0366.24
LLaVA-Video (Chain-of-Thought)Params=7B, Base Model=LLaVA-Video Base2025.0365.34
SeViLAParams=4B2025.0364.9
KTV-34B-denseLLM Size=34B, Vis Encoder=CLIP-L, Training Regime=Training-Free, Variant=dense2026.0254.7
KTV-34B-normalLLM Size=34B, Vis Encoder=CLIP-L, Training Regime=Training-Free, Variant=normal2026.0254.6
KTV-34B-sparseLLM Size=34B, Vis Encoder=CLIP-L, Training Regime=Training-Free, Variant=sparse2026.0254.2
KTV-7B-denseLLM Size=7B, Vis Encoder=CLIP-L, Training Regime=Training-Free, Variant=dense2026.0252.7
KTV-7B-normalLLM Size=7B, Vis Encoder=CLIP-L, Training Regime=Training-Free, Variant=normal2026.0252.5
KTV-7B-sparseLLM Size=7B, Vis Encoder=CLIP-L, Training Regime=Training-Free, Variant=sparse2026.0252.3
UIO-2XXLModel Size=XXL2023.1252.2
UIO-2XLModel Size=XL2023.1252
SF-LLaVA-7BLLM Size=34B, Vis Encoder=CLIP-L, Training Regime=Training-Free2026.0251.3
DYTOLLM Size=34B, Vis Encoder=CLIP-L, Training Regime=Training-Free2026.0251.1
UIO-2LModel Size=L2023.1251
DYTOLLM Size=7B, Vis Encoder=CLIP-L, Training Regime=Training-Free2026.0250.7
IG-VLMLLM Size=34B, Vis Encoder=CLIP-L, Training Regime=Training-Free2026.0250.5
SF-LLaVA-7BLLM Size=7B, Vis Encoder=CLIP-L, Training Regime=Training-Free2026.0248.8
IG-VLMLLM Size=7B, Vis Encoder=CLIP-L, Training Regime=Training-Free2026.0248.6
AnyMAL-Image 70Bzero-shot=true, vision encoder=ViT-G, LLM size=70B, number of frames=42023.0948.2
LLaMA-3.2VParams=11B2025.0345.62
TraveLERTraining Protocol=Zero-shot2024.0444.9
SeViLATraining Protocol=Uses fine-tuned components2024.0444.6
AnyMAL-Image 13Bzero-shot=true, vision encoder=ViT-G, LLM size=13B, number of frames=42023.0944.4
BLIPv2 ViTG FlanT5xxlzero-shot=true, vision encoder=ViTG, LLM=FlanT5xxl, number of frames=42023.0942.2
BLIP-2concatTraining Protocol=Zero-shot2024.0442.2
Flamingo-9Bzero-shot=true, LLM size=9B2023.0941.8
Flamingo-9BTraining Protocol=Zero-shot2024.0441.8
Internvideozero-shot=true, number of frames=82023.0941.6
InternVideoTraining Protocol=Zero-shot2024.0441.6
AnyMAL-Video 70Bzero-shot=true, vision encoder=Internvideo, LLM size=70B, number of frames=82023.0941.3
FlamingoModel Size=9B, Few-shot=true2023.1241.2
BLIP-2votingTraining Protocol=Zero-shot2024.0440.3
Flamingo-80Bzero-shot=true, LLM size=80B2023.0939.7
InstructBLIPZero-shot=true2023.1238.3
AnyMAL-Video 13Bzero-shot=true, vision encoder=Internvideo, LLM size=13B, number of frames=82023.0937.5
BLIP-2Zero-shot=true2023.1236.7
TimeChatParams=7B, Base Model=TimeChat Base2025.0321.03
TimeChat (Chain-of-Thought)Params=7B, Base Model=TimeChat Base2025.0317.54