List your company

Temporal Spatial Reasoning on TempCompass

76.8Average AccuracyG2F-RAG
58.0862.9467.872.66Apr 6, 2026

Evaluation results

MethodLinks
G2F-RAGSize=14B, Base Model=InternVL3.52026.0476.8
G2F-RAGSize=8B, Base Model=InternVL3.52026.0475.5
G2F-RAGSize=7B, Base Model=Qwen2.5-VL2026.0472
InternVL3.5Size=14B2026.0471.2
InternVL3.5Size=8B2026.0471
InternVL3Size=8B2026.0470.4
GPT-4o (2024-05-13)Size=-2026.0469.5
Qwen2.5-VLSize=7B2026.0469.2
G2F-RAGSize=7B, Base Model=LLaVA-Video2026.0469
G2F-RAGSize=4B, Base Model=InternVL3.52026.0468.7
InternVL2.5Size=8B2026.0468.7
Gemini1.5 ProSize=-2026.0468.4
LLaVA-VideoSize=7B2026.0465.5
InternVL3.5Size=4B2026.0465.4
Qwen2.5-VLSize=3B2026.0464.4
MiniCPM-V 2.6Size=8B2026.0459.6
VILA-1.5Size=8B2026.0458.8