List your company
ResearchFormerly WizWand

Cultural Multimodal Understanding on CCBench

27.9ScoreSPHINX-1k
-1.042166.4716713.985521.49933Nov 13, 2023Mar 24, 2024Aug 3, 2024Dec 14, 2024Apr 25, 2025Sep 4, 2025Jan 15, 2026
Updated 8mo ago

Evaluation results

MethodLinks
SPHINX-1kResolution=1k2023.1127.9
SPHINX-2kResolution=2k2023.1127.4
LLaVA1.5-13BModel Size=13B2023.1126.5
SPHINX2023.1125.6
LLaVA1.5-7BModel Size=7B2023.1116.4
Qwen2.5-VL2026.010.592
LLaVA-OneVision2026.010.549
Gemma3-4b-it2026.010.369
Molmo2026.010.367
PaliGemma2026.010.333
LLaVA-NeXT2026.010.243
SmolVLM22026.010.231
LLaVA-AIMv2-2D-RoPEEncoder=AIMv2, 2D-RoPE=true2026.010.114
LLaVA-2D-RoPE2D-RoPE=true2026.010.086
LLaVA-SigLIP2Encoder=SigLIP22026.010.086
LLaVA-SigLIP2-2D-RoPEEncoder=SigLIP2, 2D-RoPE=true2026.010.086
LLaVA-AIMv2Encoder=AIMv22026.010.086
LLaVA-1.52026.010.084
LLaVA-SigLIP-2D-RoPEEncoder=SigLIP, 2D-RoPE=true2026.010.08
LLaVA-SigLIPEncoder=SigLIP2026.010.071