GenAgent: Scaling Text-to-Image Generation via Agentic Multimodal Reasoning

TL;DR

GenAgent introduces an agentic multimodal framework that decouples understanding and generation, enabling autonomous multi-turn reasoning and tool invocation to improve text-to-image generation performance.

Contribution

It presents a novel agentic approach that unifies multimodal understanding and generation with reinforcement learning, allowing dynamic multi-turn interactions and improved image quality.

Findings

Achieves +23.6% on GenEval++ and +14% on WISE benchmarks.

Demonstrates cross-tool generalization and task-adaptive reasoning.

Enables test-time scaling with consistent multi-turn improvements.

Abstract

We introduce GenAgent, unifying visual understanding and generation through an agentic multimodal model. Unlike unified models that face expensive training costs and understanding-generation trade-offs, GenAgent decouples these capabilities through an agentic framework: understanding is handled by the multimodal model itself, while generation is achieved by treating image generation models as invokable tools. Crucially, unlike existing modular systems constrained by static pipelines, this design enables autonomous multi-turn interactions where the agent generates multimodal chains-of-thought encompassing reasoning, tool invocation, judgment, and reflection to iteratively refine outputs. We employ a two-stage training strategy: first, cold-start with supervised fine-tuning on high-quality tool invocation and reflection data to bootstrap agent behaviors; second, end-to-end agentic reinforcement learning combining pointwise rewards (final image quality) and pairwise rewards (reflection accuracy), with trajectory resampling for enhanced multi-turn exploration. GenAgent significantly boosts base generator(FLUX.1-dev) performance on GenEval++ (+23.6\%) and WISE (+14\%). Beyond performance gains, our framework demonstrates three key properties: 1) cross-tool generalization to generators with varying capabilities, 2) test-time scaling with consistent improvements across interaction rounds, and 3) task-adaptive reasoning that automatically adjusts to different tasks. Our code will be available at \href{https://github.com/deep-kaixun/GenAgent}{this url}.