List your company
ResearchFormerly WizWand

Multimodal Understanding on SEED Bench Img

77SEEDB ScoreQwen2.5-VL
-2.9801617.7839238.54859.31208Dec 4, 2024Feb 9, 2025Apr 18, 2025Jun 25, 2025Sep 1, 2025Nov 8, 2025Jan 15, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Qwen2.5-VL2026.0177
LLaVA-OneVision2026.0176.7
Molmo2026.0174.6
PaliGemma2026.0171.5
SmolVLM22026.0171.3
LLaVA-NeXT2026.0169.6
Gemma3-4b-it2026.0165.5
LLaVA-1.52026.0160.1
LLaVA-AIMv2Encoder=AIMv22026.0159.5
LLaVA-AIMv2-2D-RoPEEncoder=AIMv2, 2D-RoPE=true2026.0158.6
LLaVA-2D-RoPE2D-RoPE=true2026.0158.5
LLaVA-SigLIPEncoder=SigLIP2026.0156.7
LLaVA-SigLIP2Encoder=SigLIP22026.0154.8
LLaVA-SigLIP-2D-RoPEEncoder=SigLIP, 2D-RoPE=true2026.0154.5
LLaVA-SigLIP2-2D-RoPEEncoder=SigLIP2, 2D-RoPE=true2026.0152.7
TokenFlow-XLBackbone=Qwen-2.5-14B, Resolution=384, Visual Input Type=Discrete2024.120.726
ShareGPT4VBackbone=Vicuna-7B, Resolution=336, Visual Input Type=Continuous2024.120.697
ChameleonBackbone=34B from scratch, Resolution=256, Visual Input Type=Discrete2024.120.696
TokenFlow-XLBackbone=Vicuna-13B, Resolution=384, Visual Input Type=Discrete2024.120.687
EMU3Backbone=8B from scratch, Resolution=512, Visual Input Type=Discrete2024.120.682
LLaVA-1.5Backbone=Vicuna-13B, Resolution=336, Visual Input Type=Continuous2024.120.681
JanusBackbone=DeepSeek-LLM-1.3B, Resolution=384, Visual Input Type=Continuous2024.120.637
TokenFlow-LBackbone=Vicuna-13B, Resolution=256, Visual Input Type=Discrete2024.120.626
TokenFlow-BBackbone=Vicuna-13B, Resolution=224, Visual Input Type=Discrete2024.120.604
VILA-UBackbone=LLaMA-2-7B, Resolution=384, Visual Input Type=Discrete2024.120.59
InstructBLIPBackbone=Vicuna-13B, Resolution=224, Visual Input Type=Continuous2024.120.588
Qwen-VL-ChatBackbone=Qwen-7B, Resolution=448, Visual Input Type=Continuous2024.120.577
NEXT-GPTBackbone=Vicuna-7B, Resolution=224, Visual Input Type=Continuous2024.120.575
VILA-UBackbone=LLaMA-2-7B, Resolution=256, Visual Input Type=Discrete2024.120.563
SEED-LLAMABackbone=LLaMA-2-13B, Resolution=224, Visual Input Type=Discrete2024.120.537
BLIP-2Backbone=Vicuna-13B, Resolution=224, Visual Input Type=Continuous2024.120.464
LWMBackbone=LLaMA-2-7B, Resolution=256, Visual Input Type=Discrete2024.120.096