VisionCreator-R1: A Reflection-Enhanced Native Visual-Generation Agentic Model

TL;DR

VisionCreator-R1 introduces a reflection-enhanced visual generation agent with a novel training methodology, significantly improving multi-image visual content generation by addressing mid-trajectory errors through systematic reflection and plan optimization.

Contribution

It proposes a reflection-augmented agent with a new RPCO training method, enabling better correction of visual errors during multi-image generation tasks.

Findings

Outperforms Gemini2.5Pro on multiple benchmarks

Uncovers asymmetry in reflection and planning optimization in RL

Effective in both single-image and multi-image tasks

Abstract

Visual content generation has advanced from single-image to multi-image workflows, yet existing agents remain largely plan-driven and lack systematic reflection mechanisms to correct mid-trajectory visual errors. To address this limitation, we propose VisionCreator-R1, a native visual generation agent with explicit reflection, together with a Reflection-Plan Co-Optimization (RPCO) training methodology. Through extensive experiments and trajectory-level analysis, we uncover reflection-plan optimization asymmetry in reinforcement learning (RL): planning can be reliably optimized via plan rewards, while reflection learning is hindered by noisy credit assignment. Guided by this insight, our RPCO first trains on the self-constructed VCR-SFT dataset with reflection-strong single-image trajectories and planning-strong multi-image trajectories, then co-optimization on VCR-RL dataset via RL. This yields our unified VisionCreator-R1 agent, which consistently outperforms Gemini2.5Pro on existing benchmarks and our VCR-bench covering single-image and multi-image tasks.