List your company

Spatio-Temporal Grounding on V-STAR

61.9AccuracyVISD
15.82827.78939.7551.711May 7, 2026

Evaluation results

MethodLinks
VISD2026.0561.926.827.226.76.335.148.9
VisionCoach-7BModel Type=Open-source, evaluated_by_authors=true2026.0561.225.425.427.15.334.347.5
GPT-4oModel Type=Proprietary2026.0560.816.712.86.5326.838.2
Open-o3-Video-7BModel Type=Open-source, evaluated_by_authors=true2026.0560.12523.625.65.633.445.8
Gemini-2-FlashModel Type=Proprietary2026.055324.523.84.62.226.935.6
LLaVA-VideoModel Type=Open-source2026.0549.510.512.21.91.320.827.3
InternVL-2.5-8BModel Type=Open-source2026.0544.28.77.80.70.117.624.9
Video-LLaMA3Model Type=Open-source2026.0541.92323.10.90.221.727
VideoChat2Model Type=Open-source2026.0536.213.712.52.511720.3
Qwen2.5-VL-7BModel Type=Open-source, evaluated_by_authors=true2026.0533.515.413.8172.519.322.4
Oryx-1.5-7BModel Type=Open-source2026.0520.513.514.810.13.515.113.8
TRACEModel Type=Open-source2026.0517.619.117.1001213.3