Large Language Models are Universal Reasoners for Visual Generation

2026-05-05

AI Summary

This work addresses the “comprehension–generation gap” in text-to-image synthesis, where models exhibit strong language understanding yet often fail to align generated outputs with complex textual prompts. To bridge this gap, the authors propose UniReasoner, a novel framework that leverages a large language model (LLM) as a universal reasoner to first produce discrete visual sketches and then generate explicit refinement instructions through a self-evaluation mechanism. These components jointly guide a conditional diffusion model during image generation. By integrating the LLM’s reasoning and verification capabilities with discrete visual representations and a self-critique process, UniReasoner enables multimodal joint reasoning, significantly enhancing both compositional structure accuracy and semantic fidelity of the generated images while preserving high visual quality.

Abstract

Text-to-image generation has advanced rapidly with diffusion models, progressing from CLIP and T5 conditioning to unified systems where a single LLM backbone handles both visual understanding and generation. Despite the architectural unification, these systems frequently fail to faithfully align complex prompts during synthesis, even though they remain highly accurate at verifying whether an image satisfies those same prompts. We formalize this as the \emph{understanding-generation gap} and propose UniReasoner, a framework that leverages the LLM as a universal reasoner to convert its understanding strength into direct generation guidance. Given a prompt, the LLM first produces a coarse visual draft composed of discrete vision tokens. It then performs a self-critique by evaluating the draft for prompt consistency, producing a grounded textual evaluation that pinpoints what needs to be corrected. Finally, a diffusion model is conditioned jointly on the prompt, the visual draft, and the evaluation, ensuring that generation is guided by explicit corrective signals. Each signal addresses a limitation of the other: the draft provides a concrete, scene-level anchor that reduces under-specification in text-only conditioning, while the evaluation turns verification into grounded, actionable constraints that correct omissions, hallucinations, and relational errors. Experiments show that UniReasoner improves compositional alignment and semantic faithfulness under the same diffusion backbone while maintaining image quality, demonstrating a practical way to exploit LLM reasoning to close the understanding-generation gap.

Problem

• understanding-generation gap
• text-to-image generation
• large language models
• compositional alignment
• semantic faithfulness

Innovation

• understanding-generation gap
• universal reasoner
• visual draft
• self-critique
• compositional alignment