GenEval

GenEval is a benchmark used to evaluate the performance of text-to-image models by assessing their ability to follow prompts and generate corresponding images.

Papers using GenEval (4)

Sketch Then Paint: Hierarchical Reinforcement Learning for Diffusion Multi-Modal Large Language Models2026PromptEcho: Annotation-Free Reward from Vision-Language Models for Text-to-Image Reinforcement Learning2026PromptRL: Prompt Matters in RL for Flow-Based Image Generation2026Advantage Weighted Matching: Aligning RL with Pretraining in Diffusion Models2025