List your company
ResearchFormerly WizWand

Image Generation on GenEval

89Overall ScoreWan-Weaver
3750.56477.5Sep 29, 2025Nov 6, 2025Dec 15, 2025Jan 23, 2026Mar 3, 2026Apr 11, 2026May 20, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Wan-WeaverModel Category=Unified Models2026.0389
Wan-Weaver2026.0389
BAGEL-Uni-Edit (Ours)Training Stage=Stage 12026.0589
BAGEL-Uni-Edit (Ours)Training Stage=Stage 22026.0589
DuoGenModel Type=Interleaved Generation2026.0188
Bagel# Tokens=5.1T, # Params.=7B / 14B, Architecture Category=Autoregressive meets Diffusion, Prompt Rewriting (flag)=true2025.0988
BagelModel Category=Unified Models, LLM rewriter=true2026.0388
Bagel2026.0388
BAGEL-RecA2026.0588
Uni-Edit ViT (und.) 224x224Resolution=224x2242026.0588
Uni-Edit ViT (und.) 224x980Resolution=224x9802026.0588
Qwen-ImageModel Type=Generation2026.0187
BAGEL2026.0587
Uni-Edit w/o VAE dropoutVAE dropout=false2026.0587
UniPicModel Category=Unified Models2026.0386
OmniGen2Model Category=Unified Models, LLM rewriter=true2026.0386
BAGEL-AnyEdit2026.0586
Uni-Edit w/o joint trainingjoint training=false2026.0586
Uni-Edit ViT (und.) 378x980Resolution=378x9802026.0586
GPT-4o-ImageModel Type=Commercial2026.0184
Blip-3oModel Type=Unified Model2026.0184
BLIP3-oModel Category=Unified Models, LLM rewriter=true2026.0384
UniWorld-V1Model Category=Unified Models, LLM rewriter=true2026.0384
UniWorld-V12026.0384
Uni-CoTModel Type=Interleaved Generation, Interleaved generation to improve image quality=true2026.0183
X-Omni# Tokens=~1T, # Params.=10B / 20B, Architecture Category=Autoregressive meets Diffusion, Prompt Rewriting (flag)=true2025.0983
Uni-X# Tokens=240B, # Params.=3B / 4.5B, Architecture Category=Autoregressive w/o Semantic Encoder, Prompt Rewriting (flag)=true, Extended Image-Text Training (flag)=true2025.0983
FLUX.1-devModel Type=Generation2026.0182
BagelModel Type=Unified Model2026.0182
BAGEL-7BStage=Uni. 1 Stage2025.1282
Uni-X# Tokens=140B, # Params.=3B / 4.5B, Architecture Category=Autoregressive w/o Semantic Encoder, Prompt Rewriting (flag)=true2025.0982
+Uni-Edit (Ours)Base Model=Janus-Pro2026.0582
Bilp3o# Tokens=-, # Params.=4B / 9B, Architecture Category=Autoregressive meets Diffusion2025.0981
Janus-Pro-7BModel Type=Unified Model2026.0180
MetaQuery-XLModel Type=Unified Model, LLM prompt rewriting=true2026.0180
UniWorld-V1Model Type=Unified Model2026.0180
OmniGen2Model Type=Unified Model2026.0180
Janus-Pro-7BStage=Uni. 1 Stage2025.1280
Janus-Pro# Tokens=~300B, # Params.=7B, Architecture Category=Autoregressive w/ Semantic Encoder2025.0980
Janus-ProModel Category=Unified Models2026.0380
MetaQuery-XLModel Category=Unified Models, LLM rewriter=true2026.0380
Janus-Pro2026.0580
MetaQuery-LStage=Uni. 2 Stage2025.1278
Show-o2Model Category=Unified Models2026.0376
Janus-Pro-1BStage=Uni. 1 Stage2025.1273
NextStep1# Tokens=~1T, # Params.=14B, Architecture Category=Autoregressive w/ Semantic Encoder, Prompt Rewriting (flag)=true2025.0973
UnisonStage=Uni. 2 Stage2025.1270
Show-o# Tokens=~500B, # Params.=1.3B, Architecture Category=Autoregressive meets Diffusion, Semantic Alignment (flag)=true2025.0969
Show-o# Tokens=~500B, # Params.=1.3B, Architecture Category=Autoregressive meets Diffusion2025.0968
Liquid# Tokens=~90B, # Params.=7B, Architecture Category=Autoregressive w/o Semantic Encoder, Prompt Rewriting (flag)=true2025.0968
OmniGenModel Category=Unified Models2026.0368
DALLE-3Model Type=Generation2026.0167
EMU3# Tokens=-, # Params.=8B, Architecture Category=Autoregressive w/o Semantic Encoder, Prompt Rewriting (flag)=true2025.0966
FLUX.1-devModel Category=Text-to-Image Generation Models2026.0366
Emu3Model Category=Unified Models, LLM rewriter=true2026.0366
FLUX.1-dev2026.0366
MMaDAModel Type=Unified Model2026.0163
SD3-mediumModel Category=Text-to-Image Generation Models2026.0362
SDXLModel Type=Generation2026.0155
TokenFlow-13BStage=Uni. 1 Stage2025.1255
SDXLModel Category=Text-to-Image Generation Models2026.0355
Emu3Model Type=Unified Model2026.0154
Show-oModel Type=Unified Model2026.0153
Show-oStage=Uni. 1 Stage2025.1253
HART#Scales=14, FlashAttention=true, Resolution=1024x10242026.0250.9
SparVAR#Scales=14, Resolution=1024x10242026.0250.7
FastVAR#Scales=14, Resolution=1024x10242026.0250.4
LWM# Tokens=~500B, # Params.=7B, Architecture Category=Autoregressive w/o Semantic Encoder2025.0947
Chameleon# Tokens=9.2T, # Params.=34B, Architecture Category=Autoregressive w/o Semantic Encoder2025.0939