List your company

Spatial Understanding on CV-Bench 2D Overall

75.4AccuracyQwen2.5-VL
39.83249.06658.367.534Jan 15, 2026

Evaluation results

MethodLinks
Qwen2.5-VL2026.0175.4
LLaVA-OneVision2026.0173
Molmo2026.0172.8
Gemma3-4b-it2026.0165.9
PaliGemma2026.0162.4
LLaVA-NeXT2026.0160.6
SmolVLM22026.0157.7
LLaVA v1.52026.0149
LLaVA-AIMv2encoder=AIMv22026.0146.6
LLaVA-2D-RoPEpositional_embedding=2D-RoPE2026.0144.3
LLaVA-SigLIP2encoder=SigLIP22026.0144.2
LLaVA-AIMv2-2D-RoPEencoder=AIMv2, positional_embedding=2D-RoPE2026.0143.2
LLaVA-SigLIP-2D-RoPEencoder=SigLIP, positional_embedding=2D-RoPE2026.0142.5
LLaVA-SigLIP2-2D-RoPEencoder=SigLIP2, positional_embedding=2D-RoPE2026.0141.5
LLaVA-SigLIPencoder=SigLIP2026.0141.2