List your company
ResearchFormerly WizWand

Multi-modal Perception Evaluation on MME Perception

1,533Perception ScoreVILA
632.7552866.47261,100.191,333.9074Feb 7, 2026Mar 3, 2026Mar 28, 2026Apr 22, 2026May 16, 2026Jun 10, 2026Jul 5, 2026
Updated 3mo ago

Evaluation results

MethodLinks
VILAPretrained-LLM=LLaMA2-7B, Model Category=Understanding Only Model2026.021,533
LLaVA-1.5Pretrained-LLM=Vicuna-7B, Model Category=Understanding Only Model2026.021,510.7
LLaVA-1.5LLM=Vicuna-7B, Resolution=3362026.071,510.7
Qwen-VL-ChatLLM=Qwen-7B, Resolution=4482026.071,487.5
LLaVA-1.5LLM=Vicuna-7B, Resolution=3362026.071,477.4
LLaVA-SPLLM=Vicuna-7B, Resolution=3362026.071,470.7
UniTokPretrained-LLM=LLaMA-2-7B, Model Category=Unified Model2026.021,448
VanillaVision Token=576, Backbone=Qwen2.5-3B, Vision Encoder=clip-vit-large-patch14-3362026.061,446
UniWeTok-ChatPretrained-LLM=Qwen3-8B, Model Category=Unified Model2026.021,415.7
MS-ResamplerVision Token=144, Backbone=Qwen2.5-3B, Vision Encoder=clip-vit-large-patch14-3362026.061,396
LDP-v2Vision Token=144, Backbone=Qwen2.5-3B, Vision Encoder=clip-vit-large-patch14-3362026.061,394
TrustLLaVALLM=Vicuna-7B, Resolution=3362026.071,390.8
DINOv2LLM Backbone=Qwen2-7B2026.021,388.19
TokenPackerVision Token=144, Backbone=Qwen2.5-3B, Vision Encoder=clip-vit-large-patch14-3362026.061,388
Pixel-ShuffleVision Token=144, Backbone=Qwen2.5-3B, Vision Encoder=clip-vit-large-patch14-3362026.061,379
C-AbstractorVision Token=144, Backbone=Qwen2.5-3B, Vision Encoder=clip-vit-large-patch14-3362026.061,376
TokenFlow-LPretrained-LLM=Vicuna-13B, Model Category=Unified Model2026.021,365.4
VILA-UPretrained-LLM=LLaMA2-7B, Model Category=Unified Model2026.021,336.2
EVE-7B(HD)Pretrained-LLM=Vicuna-7B, Model Category=Unified Model2026.021,305.7
InternVL-ChatPretrained-LLM=Vicuna-7B, Model Category=Understanding Only Model2026.021,298.5
TrustLLaVA (MLP)LLM=Vicuna-7B, Resolution=3362026.071,296.8
mFRESALLM Backbone=Qwen2-7B2026.021,283.48
DINOv2LLM Backbone=Phi22026.021,279.21
FT-DINOSAURLLM Backbone=Qwen2-7B2026.021,242.25
StableLSDLLM Backbone=Qwen2-7B2026.021,239.48
VQDINO_MLPLLM Backbone=Qwen2-7B2026.021,229.43
DINOSAURv2LLM Backbone=Qwen2-7B2026.021,224.84
InstructBLIPPretrained-LLM=Vicuna-7B, Model Category=Understanding Only Model2026.021,212.8
mFRESALLM Backbone=Phi22026.021,187.05
DINOSAURLLM Backbone=Qwen2-7B2026.021,178.87
Slot DiffusionLLM Backbone=Qwen2-7B2026.021,171.83
SPOTLLM Backbone=Qwen2-7B2026.021,169.11
VQDINO_MLPLLM Backbone=Phi22026.021,167.08
HarmonPretrained-LLM=Qwen2.5-1.5B, Model Category=Unified Model2026.021,155
StableLSDLLM Backbone=Phi22026.021,129.08
DINOSAURv2LLM Backbone=Phi22026.021,122.73
Show-oPretrained-LLM=Phi-1.5-1.3B, Model Category=Unified Model2026.021,097.2
Slot DiffusionLLM Backbone=Phi22026.021,090.1
SPOTLLM Backbone=Phi22026.021,069.8
DINOSAURLLM Backbone=Phi22026.021,047.34
FT-DINOSAURLLM Backbone=Phi22026.021,004.94
Blind-VLMLLM Backbone=Qwen2-7B2026.02686.29
Blind-VLMLLM Backbone=Phi22026.02667.38