Paper 2602.12205
DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing
- Published
- Feb 2026
- Research lab
- Independent
- Citations
- 15
- GitHub
- 584 stars
01 In brief
Summary
DeepGen 1.0 is a lightweight 5B-parameter unified multimodal model (3B VLM + 2B DiT) for image generation and editing, achieving performance competitive with or surpassing much larger models.
It introduces Stacked Channel Bridging (SCB), which fuses features from six uniformly distributed VLM layers with learnable 'think tokens' to provide the DiT with reasoning-rich guidance.
Training uses a three-stage data-centric strategy: alignment pre-training on ~35M image-text pairs and editing triplets, joint supervised fine-tuning on a high-quality mixture of generation, editing, reasoning, and text-rendering tasks (~11M samples), and reinforcement learning with MR-GRPO, which combines multiple rewards and an auxiliary supervised diffusion loss to prevent capability degradation.
Trained on only ~50M samples, DeepGen 1.0 outperforms the 80B HunyuanImage by 28% on WISE and the 27B Qwen-Image-Edit by 37% on UniREditBench, while also achieving strong results on DPG-Bench (87.90), GenEval (0.87), and text rendering (CLIPScore 0.8278).
Ablations confirm the importance of SCB, think tokens, VLM activation, and RL components.
The model, code, and datasets are open-sourced.
02 From the paper
Abstract
Current unified multimodal models for image generation and editing typically rely on massive parameter scales (e.g., >10B), entailing prohibitive training costs and deployment footprints. In this work, we present DeepGen 1.0, a lightweight 5B unified model that achieves comprehensive capabilities competitive with or surpassing much larger counterparts. To overcome the limitations of compact models in semantic understanding and fine-grained control, we introduce Stacked Channel Bridging (SCB), a deep alignment framework that extracts hierarchical features from multiple VLM layers and fuses them with learnable 'think tokens' to provide the generative backbone with structured, reasoning-rich guidance. We further design a data-centric training strategy spanning three progressive stages: (1) Alignment Pre-training on large-scale image-text pairs and editing triplets to synchronize VLM and DiT representations, (2) Joint Supervised Fine-tuning on a high-quality mixture of generation, editing, and reasoning tasks to foster omni-capabilities, and (3) Reinforcement Learning with MR-GRPO, which leverages a mixture of reward functions and supervision signals, resulting in substantial gains in generation quality and alignment with human preferences, while maintaining stable training progress and avoiding visual artifacts. Despite being trained on only ~50M samples, DeepGen 1.0 achieves leading performance across diverse benchmarks, surpassing the 80B HunyuanImage by 28% on WISE and the 27B Qwen-Image-Edit by 37% on UniREditBench. By open-sourcing our training code, weights, and datasets, we provide an efficient, high-performance alternative to democratize unified multimodal research.