List your company

3D Question Answering on OpenEQA

56.2LLM-MatchLLaVA-OV-7B
18.55228.32638.147.874Mar 31, 2026

Evaluation results

MethodLinks
LLaVA-OV-7B2026.0356.2
GPT4-VInput Frames=50 frames2026.0355.3
MovieChatBackbone=LLaVA-OV-7B2026.0354.9
GPT4-VInput Frames=15 frames2026.0354.6
Claude-3.5 Sonnet2026.0348.7
Gemini 1.0 Pro Vision2026.0344.9
LLaMA-2Auxiliary Information=LLaVA-1.5 caption2026.0336.8
LLaMA-2Auxiliary Information=Sparse Voxel Map2026.0334.3
Video-ChatGPT2026.0332.1
AuroraCap2026.0328.9
LLaMA-2Auxiliary Information=Concept Graph2026.0328.7
Video-LLaMA2026.0320