AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

PromptRL: Prompt Matters in RL for Flow-Based Image Generation

Fu-Yun Wang, Han Zhang, Michaël Gharbi, Hongsheng Li, Taesung Park

ICML 2026 · Image generation › Text-to-image diffusion · approach: T2I evaluation and controllable adapters · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add jointly trained LM prompt refiner to flow RLFlowGRPO · GenEvaloverall score: 0.92 -> 0.97Improved
add inference-time prompt enhancement to PromptRLPromptRL · GenEvaloverall score: 0.94 -> 0.97Improved
apply naive prompt enhancement at inference to RL-tuned flow modelFlowGRPO · GenEvaloverall score: 0.92 -> 0.81Hurt
apply naive prompt enhancement at inference to RL-tuned flow modelDiffusionNFT · GenEvaloverall score: 0.88 -> 0.77Hurt
apply naive prompt enhancement (no joint training) to editing modelFLUX.1-Kontext · OmniEdit valEditReward average: 1.19 -> 1.01Hurt
add PromptRL joint training to editing modelFLUX.1-Kontext · OmniEdit valEditReward average: 1.19 -> 1.43Improved
train with 2 retained original prompts instead of 0 (ablation)PromptRL · GenEval (no PE)overall score: 0.83 -> 0.94Improved
train with 2 retained original prompts instead of 1 (ablation)PromptRL · GenEval (no PE)overall score: 0.77 -> 0.94Improved
retain 4 instead of 2 original prompts (ablation)PromptRL · GenEval (no PE)overall score: 0.94 -> 0.92Hurt
use single-reward instead of tag-based multi-reward training (ablation)PromptRL · GenEval/OCR/PickScoreGenEval: 0.93 -> 0.97Mixed
scale flow-only RL to 2x rollouts instead of using PromptRL (ablation)Flow-only RL · GenEvaloverall score: 0.97 -> 0.93Hurt
apply PromptRL PE to unseen SD3SD3 · GenEvaloverall score: 0.58 -> 0.77Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy