inclusionAI/UI-Venus-Ground-7B

VISIONPricing:Input $0.4 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:32kPublished:Aug 15, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

inclusionAI/UI-Venus-Ground-7B is a 7 billion parameter multimodal UI agent built on Qwen2.5-VL, designed for precise GUI element grounding and effective navigation using only screenshots. Developed by inclusionAI, it achieves state-of-the-art performance on UI understanding and interaction benchmarks through Reinforcement Fine-Tuning (RFT). This model excels at visual grounding, UI navigation, and complex task reasoning across various platforms.

Loading preview...

UI-Venus-Ground-7B: A High-Performance UI Agent

UI-Venus-Ground-7B is a 7 billion parameter multimodal UI agent developed by inclusionAI, based on the Qwen2.5-VL architecture. It is specifically engineered to perform accurate GUI element grounding and effective navigation using only screenshots as input, making it a powerful tool for autonomous UI interaction.

Key Capabilities & Innovations

  • State-of-the-Art Performance: Achieves leading results on critical UI benchmarks, including ScreenSpot-Pro (50.8%), ScreenSpot-v2 (94.1%), OS-World-G, and AndroidWorld (49.1% success rate).
  • Reinforcement Fine-Tuning (RFT): Utilizes carefully designed reward functions for both grounding and navigation tasks, trained on hundreds of thousands of high-quality samples.
  • Advanced Reasoning: Incorporates Self-Evolving Trajectory History Alignment and Sparse Action Enhancement to improve reasoning coherence and action distribution for better long-horizon planning.
  • Cross-Platform Generalization: Demonstrates strong capabilities in visual grounding, UI navigation, and complex task reasoning across diverse UI environments.

Ideal Use Cases

  • Automated UI Testing: For precise identification and interaction with GUI elements.
  • Robotic Process Automation (RPA): Enabling agents to navigate and operate software interfaces autonomously.
  • Accessibility Tools: Developing systems that can interpret and interact with visual interfaces for users with disabilities.
  • Research in UI Agents: Provides an open-source, high-performance baseline for advancing autonomous UI interaction.

Finetunes
1 models