Text-to-Image Generation on T2I-CompBench (Attribute Fidelity)
73.66Shape FidelityPerfGuard
Evaluation results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PerfGuard2026.01 | 73.66 | 87.53 | 81.48 | 61.2 | 37.54 | 50.07 | — | — | — | — | — | — | — | — | |
| GenArtist2026.01 | 69.48 | 84.82 | 77.09 | 54.37 | 33.46 | 44.99 | — | — | — | — | — | — | — | — | |
| LatentMorphBackbone=Janus-Pro2026.02 | 69.46 | 84.04 | 79.89 | 50.93 | 39.27 | 63.6 | 64.53 | — | — | — | — | — | — | — | |
| SD 3.5-ft-10k + MixFlowSampling steps=402025.12 | 69.12 | — | — | 35.06 | — | 45.56 | — | — | — | — | — | — | — | — | |
| SD 3.5-ft-10k + MixFlowSampling steps=102025.12 | 68.09 | — | — | 32.01 | — | 41.65 | — | — | — | — | — | — | — | — | |
| LI-DiT-10BModel=LI-DiT-10B, Parameters=10B2024.06 | 68.03 | 83.78 | 78.5 | 39.69 | — | — | — | — | — | — | — | — | — | — | |
| DALL-E 3Model=DALL-E 32024.06 | 67.5 | 81.1 | 80.7 | — | — | — | — | — | — | — | — | — | — | — | |
| DALL-E32025.03 | 67.5 | 81.1 | 80.7 | — | — | — | — | — | — | — | — | — | — | — | |
| AgentComp2026.03 | 66.81 | 87.43 | 81.42 | 47.48 | 31.96 | 42.61 | 59.62 | — | — | — | — | — | — | — | |
| SD 3.5-ft-10kSampling steps=402025.12 | 65.56 | — | — | 28.55 | — | 42.05 | — | — | — | — | — | — | — | — | |
| SD 3.5Sampling steps=402025.12 | 65.55 | — | — | 28.5 | — | 42.02 | — | — | — | — | — | — | — | — | |
| SD 3.5-ft-20kSampling steps=402025.12 | 65.52 | — | — | 28.52 | — | 42.08 | — | — | — | — | — | — | — | — | |
| SD 3.5-ft-10kSampling steps=102025.12 | 63.49 | — | — | 25.3 | — | 38.74 | — | — | — | — | — | — | — | — | |
| SD 3.5-ft-20kSampling steps=102025.12 | 63.48 | — | — | 25.32 | — | 38.78 | — | — | — | — | — | — | — | — | |
| DALLE-3+MarmotBase Model=DALLE-3, Correction Framework=Marmot, Parameter Scale=8B, Training-free status=true2025.04 | 63.48 | 83.41 | 75.73 | 30.8 | 30.55 | 38.63 | — | — | — | — | — | — | — | — | |
| SD 3.5Sampling steps=102025.12 | 63.44 | — | — | 25.1 | — | 38.72 | — | — | — | — | — | — | — | — | |
| AFS-Search-ProTime (s)=62.3, VLM supervisor=Qwen-VL-MAX2026.03 | 62.92 | 88.47 | 76.09 | 62.5 | 53.05 | 61.85 | 67.48 | — | — | — | — | — | — | — | |
| DALLE-3Base Model=DALLE-3, Correction Framework=None2025.04 | 62.13 | 80.11 | 74.64 | 22.72 | 30.42 | 37.27 | — | — | — | — | — | — | — | — | |
| DALLE·3Model Category=Diffusion Models2025.05 | 62.05 | 77.85 | 70.36 | — | 30.03 | 37.73 | — | — | — | — | — | 28.65 | — | — | |
| SMC-CFGBackbone=Flux-dev2026.03 | 61.99 | 82.16 | 79.01 | 29.39 | — | — | — | — | — | — | — | — | — | — | |
| TwiG-RLBackbone=Janus-Pro, Mode=Reinforcement Learning2026.02 | 61.28 | 82.49 | 73.19 | 34.06 | 31.99 | 54.45 | 56.24 | — | — | — | — | — | — | — | |
| AFS-Search-FastTime (s)=32.5, VLM supervisor=Qwen2.5-VL-7B2026.03 | 61.21 | 81.32 | 76.07 | 54.16 | 48.32 | 52.08 | 62.19 | — | — | — | — | — | — | — | |
| T2I-Copilot2026.01 | 61.2 | 80.39 | 76.04 | 32.28 | 33.79 | 39.85 | — | — | — | — | — | — | — | — | |
| Janus-R1Model Type=Multimodal Image Reasoning Model, reproduced=true, step=3rd2026.04 | 61.04 | 83.53 | 73.42 | 37.35 | 31.58 | 39.15 | — | — | — | — | — | — | — | — | |
| SDv3.5Time (s)=42.82026.03 | 60.5 | 77.17 | 72.5 | 22.86 | 31.76 | 37.29 | 50.35 | — | — | — | — | — | — | — | |
| Lumina-Image 2.0# Params=2.6B2025.03 | 60.28 | 82.11 | 74.12 | — | — | — | — | — | — | — | — | — | — | — | |
| SMC-CFGBackbone=SD3.52026.03 | 60.09 | 74.61 | 74.06 | 25.63 | — | — | — | — | — | — | — | — | — | — | |
| Janus-R1Model Type=Multimodal Image Reasoning Model, reproduced=true, step=2nd2026.04 | 60.04 | 84.1 | 73.39 | 37.58 | 31.33 | 38.44 | — | — | — | — | — | — | — | — | |
| Delta-KBackbone=SD3.5-M2026.03 | 59.84 | 81.25 | 78.16 | 34.87 | 33.31 | 34.1 | — | — | — | — | — | — | — | — | |
| LI-DiT-1BModel=LI-DiT-1B, Parameters=1B2024.06 | 59.34 | 74.08 | 69.59 | 27.57 | — | — | — | — | — | — | — | — | — | — | |
| SMC-CFGBackbone=Qwen-Image2026.03 | 59.34 | 81.91 | 74.21 | 40.85 | — | — | — | — | — | — | — | — | — | — | |
| SD3.5Backbone=SD3.52026.03 | 59.15 | 67.9 | 72.43 | 16.25 | — | — | — | — | — | — | — | — | — | — | |
| T2I-R1Model Type=Multimodal Image Reasoning Model, reproduced=true2026.04 | 59.14 | 82.19 | 72.66 | 35.56 | 30.93 | 40.42 | — | — | — | — | — | — | — | — | |
| Lumina-DiMOO + LADRAvg. t (s)=13.412026.03 | 58.94 | 82.25 | 72.34 | 46.35 | 31.6 | 36.19 | — | — | — | — | — | — | — | — | |
| SD32026.01 | 58.85 | 81.32 | 73.34 | 32 | 31.4 | 37.03 | — | — | — | — | — | — | — | — | |
| Stable v3Model Category=Diffusion Models2025.05 | 58.85 | 81.32 | 73.34 | — | 31.4 | 37.71 | — | — | — | — | — | 32 | — | — | |
| SD3.5 LargeModel Type=Diffusion Model2026.04 | 58.85 | 81.32 | 73.34 | 32 | 31.4 | 37.71 | — | — | — | — | — | — | — | — | |
| SD3Param=2B2026.04 | 58.64 | 80.94 | 72.97 | — | — | 37.8 | — | — | — | — | — | 32.19 | 40.44 | 60.78 | |
| T2I-R12026.01 | 58.52 | 81.3 | 72.43 | 33.78 | 30.9 | 39.93 | — | — | — | — | — | — | — | — | |
| T2I-R1Backbone=Janus-Pro2026.02 | 58.52 | 81.3 | 72.41 | 33.78 | 30.9 | 39.93 | 52.81 | — | — | — | — | — | — | — | |
| T2I-R1Time (s)=83.22026.03 | 58.52 | 81.3 | 72.43 | 33.78 | 30.9 | 39.93 | 52.81 | — | — | — | — | — | — | — | |
| NOVA (Ours)Param=0.6B2026.04 | 57.8 | 71.76 | 69.8 | — | — | 35.75 | — | — | — | — | — | 33.06 | 40.46 | 60.27 | |
| Lumina-DiMOO + ML-CacheAvg. t (s)=31.782026.03 | 57.75 | 81.52 | 70.28 | 46.79 | 31.83 | 35.23 | — | — | — | — | — | — | — | — | |
| Flux-devBackbone=Flux-dev2026.03 | 57.51 | 81.72 | 74.32 | 27.08 | — | — | — | — | — | — | — | — | — | — | |
| FLUXOrganization=Labs2026.01 | 57.18 | 74.07 | 69.22 | 28.63 | 31.27 | 37.71 | — | — | — | — | — | — | — | — | |
| FLUX.1-devParam=12B2026.04 | 57.18 | 74.07 | 69.22 | — | — | — | — | — | — | — | — | 28.63 | 38.66 | 61.85 | |
| FLUX.1Model Category=Diffusion Models2025.05 | 57.18 | 74.07 | 69.22 | — | 31.27 | 37.03 | — | — | — | — | — | 28.63 | — | — | |
| Emu3# Params=8.0B2025.03 | 57.06 | 75.44 | 71.64 | — | — | — | — | — | — | — | — | — | — | — | |
| Emu3Model Category=Auto-regressive Models2025.05 | 57.06 | 75.44 | 71.64 | — | — | — | — | — | — | — | — | — | — | — | |
| Emu3Model Type=Multimodal Large Language Model2026.04 | 57.06 | 75.44 | 71.64 | — | — | — | — | — | — | — | — | — | — | — | |
| Lumina-DiMOOAvg. t (s)=56.922026.03 | 57.02 | 81.07 | 69.3 | 46.7 | 31.7 | 34.98 | — | — | — | — | — | — | — | — | |
| NPC_SDXLbackbone=SDXL2025.12 | 56.81 | 79.5 | 68.89 | 25.45 | 31.91 | 51.71 | 52.37 | — | — | — | — | — | — | — | |
| NPCBase Model=SDXL2025.12 | 56.81 | 79.5 | 68.89 | 25.45 | 31.91 | 51.71 | 52.37 | — | — | — | — | — | — | — | |
| FiMRstep=3rd2026.04 | 56.7 | 84.74 | 75.57 | 43.05 | 31.61 | 39.35 | — | — | — | — | — | — | — | — | |
| SynGenBackbone=SD3.5-M2026.03 | 56.62 | 79.82 | 71.83 | 32.21 | 31.29 | 31.35 | — | — | — | — | — | — | — | — | |
| ConPreDiff2026.03 | 56.37 | 70.19 | 70.21 | 23.62 | 31.95 | 41.84 | 49.03 | — | — | — | — | — | — | — | |
| Uni-X# Params.=3B / 4.5B, Trained on more image-text data=true2025.09 | 56.3 | 76.5 | 67.1 | — | — | — | 66.6 | — | — | — | — | — | — | — | |
| Qwen-ImageBackbone=Qwen-Image2026.03 | 56.21 | 77.47 | 67.47 | 29.68 | — | — | — | — | — | — | — | — | — | — | |
| A&EBackbone=SD3.5-M2026.03 | 56.12 | 79.59 | 72.62 | 33.1 | 32.21 | 31.22 | — | — | — | — | — | — | — | — | |
| InitNOBackbone=SD3.5-M2026.03 | 55.91 | 80.02 | 74.99 | 32.11 | 32.4 | 32.15 | — | — | — | — | — | — | — | — | |
| PixArt-αModel=PixArt-α2024.06 | 55.82 | 68.86 | 70.44 | 20.82 | — | — | — | — | — | — | — | — | — | — | |
| PixArt-alpha# Params=0.6B2025.03 | 55.82 | 68.86 | 70.44 | — | — | — | — | — | — | — | — | — | — | — | |
| PixArt-α2025.05 | 55.82 | 68.86 | 70.44 | 20.82 | 31.79 | 41.17 | — | — | — | — | — | — | — | — | |
| PixArt-αTime (s)=6.02026.03 | 55.82 | 68.86 | 70.44 | 20.82 | 31.79 | 41.17 | 48.15 | — | — | — | — | — | — | — | |
| FiMRstep=2nd2026.04 | 55.81 | 84.17 | 75.21 | 42.92 | 31.58 | 39.15 | — | — | — | — | — | — | — | — | |
| SD3.5-MBackbone=SD3.5-M2026.03 | 55.63 | 79.39 | 73.81 | 30.53 | 31.46 | 30.5 | — | — | — | — | — | — | — | — | |
| GoT-R1-7BModel Category=Auto-regressive Models, Training Framework=GRPO, Guidance Scale=52025.05 | 55.49 | 81.39 | 73.39 | — | 31.69 | 39.44 | — | — | — | — | — | 33.06 | — | — | |
| Lumina-DiMOO + ProphetAvg. t (s)=32.432026.03 | 55.48 | 80.92 | 70.34 | 42.37 | 31.46 | 34.92 | — | — | — | — | — | — | — | — | |
| SDXL+MarmotBase Model=SDXL, Correction Framework=Marmot, Parameter Scale=8B, Training-free status=true2025.04 | 55.27 | 68.49 | 64.12 | 33.47 | 31.27 | 35.02 | — | — | — | — | — | — | — | — | |
| NOVAParam=0.6B2026.04 | 54.89 | 69.07 | 67.12 | — | — | 35.1 | — | — | — | — | — | 29.83 | 38.86 | 59.33 | |
| LMD+MarmotBase Model=LMD, Correction Framework=Marmot, Parameter Scale=8B, Training-free status=true2025.04 | 54.78 | 64.93 | 59.5 | 33.05 | 28.04 | 32.45 | — | — | — | — | — | — | — | — | |
| DALL-E 2Model=DALL-E 22024.06 | 54.64 | 57.5 | 63.74 | 12.83 | — | — | — | — | — | — | — | — | — | — | |
| DALL-E 22025.05 | 54.64 | 57.5 | 63.74 | 12.83 | 30.43 | 36.96 | — | — | — | — | — | — | — | — | |
| DALL-E 22026.03 | 54.64 | 57.5 | 63.74 | 12.83 | 30.43 | 36.96 | 42.68 | — | — | — | — | — | — | — | |
| EvoGenTime (s)=125.32026.03 | 54.57 | 71.04 | 72.34 | 21.76 | 33.08 | 42.52 | 49.22 | — | — | — | — | — | — | — | |
| Janus-R1Model Type=Multimodal Image Reasoning Model, reproduced=true, step=1st2026.04 | 54.57 | 79.25 | 70.85 | 34.18 | 31.31 | 37.74 | — | — | — | — | — | — | — | — | |
| Sana-1.5Param=1.6B2026.04 | 54.26 | 76.25 | 67.61 | — | — | 37.27 | — | — | — | — | — | 38.14 | 40.88 | 61.1 | |
| LMD+GenArtistBase Model=LMD, Correction Framework=GenArtist2025.04 | 54.18 | 56.4 | 59.56 | 32.45 | 28.25 | 32 | — | — | — | — | — | — | — | — | |
| SDXL# Params.=3.5B2025.09 | 54.1 | 63.7 | 56.4 | — | — | — | 58.1 | — | — | — | — | — | — | — | |
| SD XLModel=SD XL2024.06 | 54.08 | 63.69 | 56.37 | 20.32 | — | — | — | — | — | — | — | — | — | — | |
| SDXL# Params=2.6B2025.03 | 54.08 | 63.69 | 56.37 | — | — | — | — | — | — | — | — | — | — | — | |
| SDXL2025.12 | 54.08 | 63.69 | 56.37 | 20.32 | 31.1 | 40.91 | 44.41 | — | — | — | — | — | — | — | |
| SDXLTime (s)=4.62026.03 | 54.08 | 63.69 | 56.37 | 20.32 | 31.79 | 40.91 | 44.53 | — | — | — | — | — | — | — | |
| Qwen-ImageTime (s)=32.22026.03 | 54.01 | 78.35 | 68.16 | 36.47 | 31.09 | 35.3 | 50.56 | — | — | — | — | — | — | — | |
| NOVA + GRPOParam=0.6B2026.04 | 53.92 | 67.21 | 66.9 | — | — | 34.84 | — | — | — | — | — | 30.69 | 39.84 | 59.7 | |
| CoMat-SDXLbase_model=SDXL2024.04 | 53.29 | 78.27 | 64.68 | 24.28 | 31.87 | 36.8 | — | — | — | — | — | — | — | — | |
| CoMatapproach_type=feedback-based2025.12 | 53.29 | 78.27 | 64.68 | 24.28 | 31.87 | 36.8 | 48.19 | — | — | — | — | — | — | — | |
| FiMRstep=1st2026.04 | 53.29 | 81.04 | 73.83 | 42.97 | 31.62 | 38.29 | — | — | — | — | — | — | — | — | |
| Show-oParam=1.3B2026.04 | 52.64 | 73.27 | 68.15 | — | — | 35.72 | — | — | — | — | — | 36.97 | 39.96 | 62.09 | |
| Liquid# Params.=7B2025.09 | 52.3 | 71.5 | 65.1 | — | — | — | 63 | — | — | — | — | — | — | — | |
| UnitokModel Type=Multimodal Large Language Model2026.04 | 51.95 | 77.45 | 64.31 | 26.68 | 31.31 | 38.3 | — | — | — | — | — | — | — | — | |
| UniToken# Params.=7B2025.09 | 51.8 | 71.2 | 66.7 | — | — | — | 63.2 | — | — | — | — | — | — | — | |
| GoT-R1-1BModel Category=Auto-regressive Models, Training Framework=GRPO, Guidance Scale=52025.05 | 51.74 | 76.32 | 65.89 | — | 31.01 | 37.49 | — | — | — | — | — | 26.74 | — | — | |
| Delta-KBackbone=SDXL2026.03 | 51.19 | 63.71 | 62.18 | 24.66 | 31.75 | 35.32 | — | — | — | — | — | — | — | — | |
| MILRBackbone=Janus-Pro2026.02 | 51.17 | 85.08 | 69.49 | 46.13 | 30.78 | 36.84 | 53.25 | — | — | — | — | — | — | — | |
| FLUXTime (s)=11.72026.03 | 51.12 | 77.36 | 63.25 | 27.47 | 30.77 | 36.22 | 47.7 | — | — | — | — | — | — | — | |
| LMDBase Model=LMD, Correction Framework=None2025.04 | 51.04 | 51.5 | 56.9 | 27.72 | 27.64 | 30.33 | — | — | — | — | — | — | — | — | |
| Playground-v22024.04 | 50.87 | 62.08 | 61.25 | 23.72 | 30.98 | 36.13 | — | — | — | — | — | — | — | — | |
| Playground-v22026.03 | 50.87 | 62.08 | 61.25 | 23.72 | 30.98 | 36.13 | — | — | — | — | — | — | — | — | |
| Flux.1 DevModel Type=Diffusion Model2026.04 | 50.66 | 75.72 | 63 | 27 | 30.65 | 36.28 | — | — | — | — | — | — | — | — |