Skywork UniPic: Unified Autoregressive Modeling for Visual Understanding and Generation

TL;DR

Skywork UniPic presents a 1.5B-parameter unified autoregressive model that integrates image understanding, generation, and editing in a single architecture through decoupled visual encoders feeding a shared LLM. It combines MAR-based generation with SigLIP2-based understanding, and trains with a four-stage curriculum and reward-driven data quality pipelines to achieve competitive results at high resolutions on commodity hardware. Key contributions include the decoupled encoding strategy, a progressive training schedule, and curated reward-based data quality assurance that enable state-of-the-art-like performance across GenEval, DPG-Bench, and editing benchmarks while maintaining efficiency. The work demonstrates that high-fidelity, versatile multimodal capabilities can be realized without large-scale parameter growth, offering a practical path to deployable multimodal AI in resource-constrained settings.

Abstract

We introduce Skywork UniPic, a 1.5 billion-parameter autoregressive model that unifies image understanding, text-to-image generation, and image editing within a single architecture-eliminating the need for task-specific adapters or inter-module connectors-and demonstrate that compact multimodal systems can achieve state-of-the-art performance on commodity hardware. Skywork UniPic achieves a GenEval score of 0.86, surpassing most existing unified models; sets a new DPG-Bench complex-generation record of 85.5; attains 5.83 on GEditBench-EN and 3.49 on ImgEdit-Bench for image editing; and generates 1024 x 1024 images with under 15 GB of GPU memory (e.g., RTX 4090). (1) a decoupled encoding strategy that leverages a masked autoregressive encoder for synthesis and a SigLIP2 encoder for understanding, all feeding a shared autoregressive decoder; (2) a progressive, resolution-aware training schedule scaling from 256 x 256 to 1024 x 1024 while dynamically unfreezing parameters to balance capacity and stability; and (3) meticulously curated, 100 million-scale datasets augmented with task-specific reward models to refine generation and editing objectives. By demonstrating that high-fidelity multimodal integration need not incur prohibitive resource demands, Skywork UniPic establishes a practical paradigm for deployable, high-fidelity multimodal AI. Code and weights are publicly available at https://huggingface.co/Skywork/Skywork-UniPic-1.5B.

Figures (6)

• Figure 1: Showcases of our model's performance on editing and generation tasks.
• Figure 2: The overall framework of Skywork UniPic. (a) Image generation is achieved through a masked auto-regressive process using the MAR model li2024autoregressiveimagegenerationvector. (b) Image understanding is performed using a SigLIP2 encoder tschannen2025siglip2multilingualvisionlanguage to extract rich visual features, which are subsequently passed to an LLM for autoregressive text generation. They share a single LLM to promote consistent instruction-following and enable knowledge transfer between generation and understanding tasks
• Figure 3: Performance comparison across multiple benchmarks. Skywork UniPic demonstrates competitive performance across understanding, generation, editing, and in-context tasks while maintaining exceptional parameter efficiency with only 1.5B activated parameters.
• Figure 4: Qualitative comparison of text-to-image generation results. Skywork UniPic produces high-quality images that accurately reflect textual prompts while maintaining competitive visual fidelity compared to both open-source and proprietary models.
• Figure 5: Qualitative comparison of image editing results. Skywork UniPic successfully handles diverse editing instructions while preserving image quality and maintaining consistency in unmodified regions, demonstrating the effectiveness of our unified approach.