Unified Thinker: A General Reasoning Modular Core for Image Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911564101058560 |
|---|---|
| author | Zhou, Sashuai Zhou, Qiang Hu, Jijin Yang, Hanqing Cao, Yue Ma, Junpeng Ma, Yinchao Song, Jun Ge, Tiezheng Yu, Cheng Zheng, Bo Zhao, Zhou |
| author_facet | Zhou, Sashuai Zhou, Qiang Hu, Jijin Yang, Hanqing Cao, Yue Ma, Junpeng Ma, Yinchao Song, Jun Ge, Tiezheng Yu, Cheng Zheng, Bo Zhao, Zhou |
| contents | Despite impressive progress in high-fidelity image synthesis, generative models still struggle with logic-intensive instruction following, exposing a persistent reasoning--execution gap. Meanwhile, closed-source systems (e.g., Nano Banana) have demonstrated strong reasoning-driven image generation, highlighting a substantial gap to current open-source models. We argue that closing this gap requires not merely better visual generators, but executable reasoning: decomposing high-level intents into grounded, verifiable plans that directly steer the generative process. To this end, we propose Unified Thinker, a task-agnostic reasoning architecture for general image generation, designed as a unified planning core that can plug into diverse generators and workflows. Unified Thinker decouples a dedicated Thinker from the image Generator, enabling modular upgrades of reasoning without retraining the entire generative model. We further introduce a two-stage training paradigm: we first build a structured planning interface for the Thinker, then apply reinforcement learning to ground its policy in pixel-level feedback, encouraging plans that optimize visual correctness over textual plausibility. Extensive experiments on text-to-image generation and image editing show that Unified Thinker substantially improves image reasoning and generation quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_03127 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Unified Thinker: A General Reasoning Modular Core for Image Generation Zhou, Sashuai Zhou, Qiang Hu, Jijin Yang, Hanqing Cao, Yue Ma, Junpeng Ma, Yinchao Song, Jun Ge, Tiezheng Yu, Cheng Zheng, Bo Zhao, Zhou Computer Vision and Pattern Recognition Artificial Intelligence Despite impressive progress in high-fidelity image synthesis, generative models still struggle with logic-intensive instruction following, exposing a persistent reasoning--execution gap. Meanwhile, closed-source systems (e.g., Nano Banana) have demonstrated strong reasoning-driven image generation, highlighting a substantial gap to current open-source models. We argue that closing this gap requires not merely better visual generators, but executable reasoning: decomposing high-level intents into grounded, verifiable plans that directly steer the generative process. To this end, we propose Unified Thinker, a task-agnostic reasoning architecture for general image generation, designed as a unified planning core that can plug into diverse generators and workflows. Unified Thinker decouples a dedicated Thinker from the image Generator, enabling modular upgrades of reasoning without retraining the entire generative model. We further introduce a two-stage training paradigm: we first build a structured planning interface for the Thinker, then apply reinforcement learning to ground its policy in pixel-level feedback, encouraging plans that optimize visual correctness over textual plausibility. Extensive experiments on text-to-image generation and image editing show that Unified Thinker substantially improves image reasoning and generation quality. |
| title | Unified Thinker: A General Reasoning Modular Core for Image Generation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2601.03127 |