UniReason 1.0: A Unified Reasoning Framework for World Knowledge Aligned Image Generation and Editing

TL;DR

UniReason tackles the limitation of current unified multimodal models by embedding world-knowledge reasoning into the generation process and coupling it with fine-grained, editing-like visual refinement. It introduces two complementary paradigms—World Knowledge-Enhanced Textual Reasoning and Fine-grained Editing-like Visual Refinement—and a two-stage training strategy to jointly learn understanding, generation, and refinement within a shared framework. A large-scale, reasoning-centric dataset across five knowledge domains supports planning, while an agent-based refinement pipeline produces supervision data for iterative improvement. Experimental results on WISE, KrisBench, UniREditBench, GenEval, and DPGBench demonstrate strong reasoning performance without sacrificing general synthesis and editing capabilities, highlighting the practical potential of unified, knowledge-grounded multimodal reasoning.

Abstract

Unified multimodal models often struggle with complex synthesis tasks that demand deep reasoning, and typically treat text-to-image generation and image editing as isolated capabilities rather than interconnected reasoning steps. To address this, we propose UniReason, a unified framework that harmonizes these two tasks through a dual reasoning paradigm. We formulate generation as world knowledge-enhanced planning to inject implicit constraints, and leverage editing capabilities for fine-grained visual refinement to further correct visual errors via self-reflection. This approach unifies generation and editing within a shared representation, mirroring the human cognitive process of planning followed by refinement. We support this framework by systematically constructing a large-scale reasoning-centric dataset (~300k samples) covering five major knowledge domains (e.g., cultural commonsense, physics, etc.) for planning, alongside an agent-generated corpus for visual self-correction. Extensive experiments demonstrate that UniReason achieves advanced performance on reasoning-intensive benchmarks such as WISE, KrisBench and UniREditBench, while maintaining superior general synthesis capabilities.

Figures (5)

• Figure 1: Illustrative cases of UniReason on image editing and T2I generation tasks. Given an instruction, the model first performs world knowledge-enhanced textual reasoning to generate grounded, fine-grained guidance for image synthesis. It then applies fine-grained editing-like visual refinement, correcting errors introduced during the initial generation and improving the synthesis quality.
• Figure 2: Overview of UniReason framework for two complementary reasoning paradigms in image synthesis.
• Figure 3: Correlation between image editing capability (ImgEdit score) and performance gains from refinement across three benchmarks. Higher editing proficiency leads to monotonically increasing refinement effectiveness.
• Figure 4: Overview of our data preparation framework.
• Figure 5: Qualitative results of UniReason on both T2I generation (blue column) and image editing task (orange column).