Skywork

Skywork-AI

Welcome to Skywork! We are an innovative research team dedicated to advancing LLM and multimodal understanding. Our mission is to push the boundaries of AI by developing cutting-edge models and datasets that enable seamless interaction between vision and language.
The name “Skywork” combines the limitless potential of the Sky with the impactful outcomes of our Work.

Models

• Skywork-R1V4: Toward Agentic Multimodal Intelligence through Interleaved Thinking with Images and DeepResearch

• Skywork-UniPic2: Building Kontext Model with Online RL for Unified Multimodal Model

• Skywork-UniPic: Unified Autoregressive Modeling for Visual Understanding and Generation

• Skywork-R1V3-38B: Advanced Multimodal Reasoning Model

• Skywork-Reward-V2 series: Scaling Preferenec Data Curation to the Extreme

• Skywork-SWE-32B: Unveiling Data Scaling Laws for Software Engineering in LLMs

• Skywork-OR1: SOTA 7B and 32B Open Reasoning Models for Math and Code

• Skywork-R1V2-38B: Multimodal Hybrid Reinforcement Learning for Reasoning

• Skywork-VL-Reward-7B: MultiModal Reward Model

• Skywork-R1V-38B: Pioneering Multimodal Reasoning with Chain-of-Thought

• SkyReels-V1: Human-Centric Video Foundation Model

• SkyReels-AX: Controllable Video Generation Series

• Skywork-o1-Open-PRM: o1 Open Model Series

• Skywork-Reward series: Strong 8B and 27B Reward Models Topping RewardBench