Text-to-Image Generation on GenEval (Compositional Metrics including IAA, IQA, ISTA)
94Overall ScoreRAISE
Evaluation results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RAISEAvg. #Samples Generated=18.6, Avg. #Calls VLM=7.32026.02 | 94 | 100 | 100 | 95 | 98 | 83 | 87 | — | — | — | — | — | — | |
| RAISEAvg. #Samples Generated=18.6, Avg. #Calls VLM=7.32026.02 | 94 | 100 | 100 | 95 | 98 | 83 | 87 | — | — | — | — | — | — | |
| Lumina-DiMOOSteps (T)=32, Trajectory (N)=32, K=8, Search Strategy=HTS2025.12 | 92 | 100 | 97 | 91 | 94 | 89 | 82 | — | — | — | — | — | — | |
| Lumina-DiMOOSteps (T)=32, Trajectory (N)=32, Search Strategy=LTS2025.12 | 91 | 100 | 97 | 89 | 93 | 87 | 80 | — | — | — | — | — | — | |
| Qwen-Image-RLAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 91 | 100 | 95 | 93 | 92 | 87 | 83 | — | — | — | — | — | — | |
| FLUX.1-dev + ReflectionFlowAvg. #Samples Generated=32, Avg. #Calls VLM=642026.02 | 91 | 100 | 98 | 89 | 95 | 89 | 75 | — | — | — | — | — | — | |
| Qwen-Image-RLAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 91 | 100 | 95 | 93 | 92 | 87 | 83 | — | — | — | — | — | — | |
| FLUX.1-dev + ReflectionFlowAvg. #Samples Generated=32, Avg. #Calls VLM=642026.02 | 91 | 100 | 98 | 89 | 95 | 89 | 75 | — | — | — | — | — | — | |
| Lumina-DiMOOSteps (T)=16, Trajectory (N)=16, K=4, Search Strategy=HTS2025.12 | 89 | 100 | 94 | 84 | 92 | 86 | 79 | — | — | — | — | — | — | |
| MogaoModel category=Unified Multimodal Models2026.03 | 89 | 100 | 97 | 83 | 93 | 84 | — | — | — | — | — | 80 | — | |
| IOMM-B 512Resolution=512, Backbone Scale=1.6B parameters2026.03 | 89 | 99 | 92 | 83 | 94 | 91 | — | — | — | — | — | 75 | — | |
| Lumina-DiMOOResolution=1024x10242025.12 | 88 | 100 | 94 | 85 | 89 | 85 | 76 | 51.93 | 71.98 | 30.86 | 51.59 | — | — | |
| BAGEL + RewriterAvg. #Samples Generated=1, Avg. #Calls VLM=22026.02 | 88 | 98 | 95 | 84 | 95 | 78 | 77 | — | — | — | — | — | — | |
| BAGEL + RewriterAvg. #Samples Generated=1, Avg. #Calls VLM=22026.02 | 88 | 98 | 95 | 84 | 95 | 78 | 77 | — | — | — | — | — | — | |
| BAGELModel category=Unified Multimodal Models, LLM rewriters=true2026.03 | 88 | 98 | 95 | 84 | 95 | 78 | — | — | — | — | — | 77 | — | |
| BAGEL-7BCategory=Unified Models, Prompt Strategy=Rewritten prompts2026.03 | 88 | 98 | 95 | 84 | 95 | 78 | — | — | — | — | — | 77 | — | |
| Qwen-ImageResolution=1024x10242025.12 | 87 | 99 | 92 | 89 | 88 | 76 | 77 | 52.02 | 74.44 | 34.13 | 53.53 | — | — | |
| Qwen-Image2025.12 | 87 | 99 | 92 | 89 | 88 | 76 | 77 | — | — | — | — | — | — | |
| Lumina-DiMOOSteps (T)=16, Trajectory (N)=16, Search Strategy=LTS2025.12 | 87 | 100 | 93 | 81 | 90 | 83 | 76 | — | — | — | — | — | — | |
| Qwen-ImageAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 87 | 99 | 92 | 89 | 88 | 76 | 77 | — | — | — | — | — | — | |
| FLUX.1-dev + Noise & Prompt ScalingAvg. #Samples Generated=32, Avg. #Calls VLM=322026.02 | 87 | 99 | 94 | 85 | 91 | 80 | 71 | — | — | — | — | — | — | |
| Qwen-ImageAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 87 | 99 | 92 | 89 | 88 | 76 | 77 | — | — | — | — | — | — | |
| FLUX.1-dev + Noise & Prompt ScalingAvg. #Samples Generated=32, Avg. #Calls VLM=322026.02 | 87 | 99 | 94 | 85 | 91 | 80 | 71 | — | — | — | — | — | — | |
| Qwen-Image#Params=20B2026.03 | 87 | 99 | 92 | 89 | 88 | 76 | — | — | — | — | — | 77 | — | |
| UniComModel category=Unified Multimodal Models2026.03 | 87 | 98 | 94 | 81 | 91 | 82 | — | — | — | — | — | 77 | — | |
| IOMM-B 1024Resolution=1024, Backbone Scale=1.6B parameters2026.03 | 87 | 99 | 91 | 75 | 93 | 88 | — | — | — | — | — | 75 | — | |
| IOMM-L 512Resolution=512, Backbone Scale=2.7B parameters2026.03 | 87 | 99 | 91 | 82 | 94 | 85 | — | — | — | — | — | 72 | — | |
| OmniGen2Resolution=1024x10242025.12 | 86 | 99 | 96 | 74 | 98 | 71 | 75 | 54.2 | 75.16 | 34.48 | 54.61 | — | — | |
| OmniGen2Model category=Unified Multimodal Models, LLM rewriters=true2026.03 | 86 | 99 | 96 | 74 | 98 | 71 | — | — | — | — | — | 75 | — | |
| FLUX.1-dev + Noise ScalingAvg. #Samples Generated=32, Avg. #Calls VLM=02026.02 | 85 | 100 | 96 | 91 | 91 | 52 | 78 | — | — | — | — | — | — | |
| FLUX.1-dev + Noise ScalingAvg. #Samples Generated=32, Avg. #Calls VLM=02026.02 | 85 | 100 | 96 | 91 | 91 | 52 | 78 | — | — | — | — | — | — | |
| InternVL-U#Params=2B+1.7B2026.03 | 85 | 99 | 94 | 74 | 91 | 77 | — | — | — | — | — | 74 | — | |
| Ming-UniVisionModel category=Unified Multimodal Models2026.03 | 85 | 100 | 93 | 59 | 93 | 92 | — | — | — | — | — | 70 | — | |
| IOMM-L 1024Resolution=1024, Backbone Scale=2.7B parameters2026.03 | 85 | 100 | 91 | 71 | 92 | 78 | — | — | — | — | — | 78 | — | |
| GPT-Image-1Resolution=Variable (not controlled)2025.12 | 84 | 99 | 92 | 85 | 92 | 75 | 61 | 69.07 | 76.74 | 51.26 | 65.69 | — | — | |
| VACoTType=Unified2025.12 | 84 | 99 | 95 | 80 | 90 | 66 | 71 | — | — | — | — | — | — | |
| GPT-4o2025.12 | 84 | 99 | 92 | 85 | 92 | 75 | 61 | — | — | — | — | — | — | |
| Seedream 3.0Avg. #Samples Generated=1, Avg. #Calls VLM=02026.02 | 84 | 99 | 96 | 91 | 93 | 47 | 80 | — | — | — | — | — | — | |
| GPT Image 1 [High]Avg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 84 | 99 | 92 | 85 | 92 | 75 | 61 | — | — | — | — | — | — | |
| Seedream 3.0Avg. #Samples Generated=1, Avg. #Calls VLM=02026.02 | 84 | 99 | 96 | 91 | 93 | 47 | 80 | — | — | — | — | — | — | |
| GPT Image 1 [High]Avg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 84 | 99 | 92 | 85 | 92 | 75 | 61 | — | — | — | — | — | — | |
| Seedream 3.0#Params=-2026.03 | 84 | 99 | 96 | 91 | 93 | 47 | — | — | — | — | — | 80 | — | |
| GPT Image 1 [High]#Params=-2026.03 | 84 | 99 | 92 | 85 | 92 | 75 | — | — | — | — | — | 61 | — | |
| Z-Image#Params=6B2026.03 | 84 | 100 | 94 | 78 | 93 | 62 | — | — | — | — | — | 77 | — | |
| TarModel category=Unified Multimodal Models2026.03 | 84 | 99 | 92 | 83 | 85 | 80 | — | — | — | — | — | 65 | — | |
| BLIP3-oModel category=Unified Multimodal Models2026.03 | 84 | — | — | — | — | — | — | — | — | — | — | — | — | |
| UniWorld-V1Model category=Unified Multimodal Models, LLM rewriters=true2026.03 | 84 | 98 | 93 | 81 | 89 | 74 | — | — | — | — | — | 71 | — | |
| BLIP3-o-8B*Category=Unified Models, Training Data=Additional 30M proprietary image-text pairs2026.03 | 84 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Uni-CoTType=Unified2025.12 | 83 | 99 | 95 | 82 | 89 | 60 | 72 | — | — | — | — | — | — | |
| HiDream-I1-FullAvg. #Samples Generated=1, Avg. #Calls VLM=02026.02 | 83 | 100 | 98 | 79 | 91 | 60 | 72 | — | — | — | — | — | — | |
| HiDream-I1-FullAvg. #Samples Generated=1, Avg. #Calls VLM=02026.02 | 83 | 100 | 98 | 79 | 91 | 60 | 72 | — | — | — | — | — | — | |
| X-OmniModel category=Unified Multimodal Models2026.03 | 83 | 98 | 95 | 75 | 91 | 71 | — | — | — | — | — | 68 | — | |
| BAGELResolution=1024x10242025.12 | 82 | 99 | 94 | 81 | 88 | 64 | 63 | 58.68 | 71.24 | 38.35 | 56.09 | — | — | |
| FLUX.1-devType=Gen. Only2025.12 | 82 | 98 | 93 | 75 | 93 | 68 | 65 | — | — | — | — | — | — | |
| UiGType=Unified2025.12 | 82 | 99 | 92 | 81 | 89 | 61 | 69 | — | — | — | — | — | — | |
| BAGEL2025.12 | 82 | 99 | 94 | 81 | 88 | 64 | 63 | — | — | — | — | — | — | |
| BAGELAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 82 | 99 | 94 | 81 | 88 | 64 | 63 | — | — | — | — | — | — | |
| BAGELAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 82 | 99 | 94 | 81 | 88 | 64 | 63 | — | — | — | — | — | — | |
| BAGEL#Params=7B+7B2026.03 | 82 | 99 | 94 | 81 | 88 | 64 | — | — | — | — | — | 63 | — | |
| FLUX.1 [Dev]Model category=Generation-only Models2026.03 | 82 | 98 | 93 | 75 | 93 | 68 | — | — | — | — | — | 65 | — | |
| SANA-1.0-1.6B + Reflect-DiTAvg. #Samples Generated=≤20, Avg. #Calls VLM=≤202026.02 | 81 | 98 | 96 | 80 | 88 | 66 | 60 | — | — | — | — | — | — | |
| SANA-1.0-1.6B + Reflect-DiTAvg. #Samples Generated=≤20, Avg. #Calls VLM=≤202026.02 | 81 | 98 | 96 | 80 | 88 | 66 | 60 | — | — | — | — | — | — | |
| BLIP3-o-4BCategory=Unified Models2026.03 | 81 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Janus-Pro-7BType=Unified2025.12 | 80 | 99 | 89 | 59 | 90 | 79 | 66 | — | — | — | — | — | — | |
| MetaQuery-XLType=Unified2025.12 | 80 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Janus-Pro2025.12 | 80 | 99 | 89 | 59 | 90 | 79 | 66 | — | — | — | — | — | — | |
| Janus-Pro-7BAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 80 | 99 | 89 | 59 | 90 | 79 | 66 | — | — | — | — | — | — | |
| Janus-Pro-7BAvg. #Samples Generated=1, Avg. #Calls VLM=12026.02 | 80 | 99 | 89 | 59 | 90 | 79 | 66 | — | — | — | — | — | — | |
| SANA-1.0-1.6B + Noise ScalingAvg. #Samples Generated=20, Avg. #Calls VLM=02026.02 | 80 | 100 | 93 | 79 | 91 | 55 | 62 | — | — | — | — | — | — | |
| Janus-Pro#Params=7B2026.03 | 80 | 99 | 89 | 59 | 90 | 79 | — | — | — | — | — | 66 | — | |
| UniWorld-V1#Params=7B+13B2026.03 | 80 | 99 | 93 | 79 | 89 | 49 | — | — | — | — | — | 70 | — | |
| OmniGen2#Params=3B+4B2026.03 | 80 | 100 | 95 | 64 | 88 | 55 | — | — | — | — | — | 76 | — | |
| MetaQuery-XLModel category=Unified Multimodal Models, LLM rewriters=true2026.03 | 80 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Janus-ProModel category=Unified Multimodal Models2026.03 | 80 | 99 | 89 | 59 | 90 | 79 | — | — | — | — | — | 66 | — | |
| Janus-Pro-7BCategory=Unified Models2026.03 | 80 | 99 | 89 | 59 | 90 | 79 | — | — | — | — | — | 66 | — | |
| MetaQuery-XLCategory=Unified Models, Prompt Strategy=Rewritten prompts2026.03 | 80 | — | — | — | — | — | — | — | — | — | — | — | — | |
| BAGELType=Unified2025.12 | 79 | 99 | 92 | 78 | 87 | 53 | 64 | — | — | — | — | — | — | |
| Infinity-8B#Speed=1.0×, #Latency=4.80s, #Param=8.0B2025.12 | 79 | — | 90 | — | 67 | 62 | — | — | — | — | — | — | — | |
| Lumina-DiMOOSteps (T)=8, Trajectory (N)=12025.12 | 78 | 96 | 85 | 70 | 85 | 72 | 62 | — | — | — | — | — | — | |
| StageVARBase Model=Infinity-8B, #Speed=2.7×, #Latency=1.77s, #Param=8.0B2025.12 | 78 | — | 87 | — | 66 | 60 | — | — | — | — | — | — | — | |
| MetaQuery-LCategory=Unified Models, Prompt Strategy=Rewritten prompts2026.03 | 78 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Meissonic + MaskFocusCategory=Masked Generative Model2025.12 | 76 | 100 | 91 | 85 | 87 | 42 | 54 | — | — | — | — | — | — | |
| Playground v3Avg. #Samples Generated=1, Avg. #Calls VLM=02026.02 | 76 | 99 | 95 | 72 | 82 | 50 | 54 | — | — | — | — | — | — | |
| Playground v3Avg. #Samples Generated=1, Avg. #Calls VLM=02026.02 | 76 | 99 | 95 | 72 | 82 | 50 | 54 | — | — | — | — | — | — | |
| Show-o2#Params=7B2026.03 | 76 | 100 | 87 | 58 | 92 | 52 | — | — | — | — | — | 62 | — | |
| HarmonModel category=Unified Multimodal Models2026.03 | 76 | 99 | 86 | 66 | 85 | 74 | — | — | — | — | — | 48 | — | |
| Show-o2Model category=Unified Multimodal Models2026.03 | 76 | 100 | 87 | 58 | 92 | 52 | — | — | — | — | — | 62 | — | |
| Show-o2-7BCategory=Unified Models, Prompt Strategy=Rewritten prompts2026.03 | 76 | 100 | 87 | 58 | 92 | 52 | — | — | — | — | — | 62 | — | |
| RPiAEDDT Head usage=with DDT Head2026.03 | 76 | 97 | 86 | 62 | 89 | 66 | — | — | — | — | — | 56 | — | |
| RPiAEDDT Head usage=not included2026.03 | 75 | 96 | 86 | 64 | 82 | 63 | — | — | — | — | — | 58 | — | |
| SD3-MediumType=Gen. Only2025.12 | 74 | 99 | 94 | 72 | 89 | 33 | 60 | — | — | — | — | — | — | |
| FLUX.1-dev + T2I-CopilotAvg. #Samples Generated=11.3, Avg. #Calls VLM=22.62026.02 | 74 | 99 | 91 | 68 | 86 | 55 | 46 | — | — | — | — | — | — | |
| FLUX.1-dev + T2I-CopilotAvg. #Samples Generated=11.3, Avg. #Calls VLM=22.62026.02 | 74 | 99 | 91 | 68 | 86 | 55 | 46 | — | — | — | — | — | — | |
| SD3-Medium#Params=2B2026.03 | 74 | 99 | 94 | 72 | 89 | 33 | — | — | — | — | — | 60 | — | |
| SD3-MediumModel category=Generation-only Models2026.03 | 74 | 99 | 94 | 72 | 89 | 33 | — | — | — | — | — | 60 | — | |
| SD3-MediumCategory=Gen. Only2026.03 | 74 | 99 | 94 | 72 | 89 | 33 | — | — | — | — | — | 60 | — | |
| MetaQuery-BCategory=Unified Models, Prompt Strategy=Rewritten prompts2026.03 | 74 | — | — | — | — | — | — | — | — | — | — | — | — | |
| Janus-Pro-1BCategory=AR-based method2025.12 | 73 | 98 | 82 | 51 | 89 | 65 | 56 | — | — | — | — | — | — | |
| Mask-GRPOCategory=Masked Generative Model2025.12 | 73 | 99 | 90 | 69 | 85 | 35 | 59 | — | — | — | — | — | — | |
| Meissonic + MaskGRPOCategory=Masked Generative Model2025.12 | 73 | 100 | 87 | 83 | 87 | 39 | 48 | — | — | — | — | — | — |