FLUX-Reason-6M & PRISM-Bench: A Million-Scale Text-to-Image Reasoning Dataset and Comprehensive Benchmark

Benchmark Model Rank Results
text-to-image-generation-on-prism-benchQwen-Image#3Accuracy: 74.1
text-to-image-generation-on-prism-benchFLUX.1-dev#4Accuracy: 68.5
text-to-image-generation-on-prism-benchSD3-Medium#5Accuracy: 66.1
text-to-image-generation-on-prism-benchSDXL#7Accuracy: 57.0