Hcompany/Holo1.5-7B
Hcompany/Holo1.5-7B is a 7 billion parameter Vision-Language Model (VLM) developed by H Company, specifically designed as a foundational model for Computer Use (CU) agents. It excels at user interface (UI) localization and UI-based question answering across web, desktop, and mobile environments, supporting high-resolution inputs up to 3840x2160 pixels. This model achieves state-of-the-art performance in UI localization and screen content understanding, outperforming previous models by an average of 4.5% in localization accuracy and 3.9% in GUI QA tasks.
Loading preview...
Holo1.5-7B: Foundational Model for Computer Use Agents
Holo1.5-7B is a 7 billion parameter Vision-Language Model (VLM) developed by H Company, specifically engineered to power Computer Use (CU) agents. These agents interact with real applications (web, desktop, mobile) by navigating interfaces, manipulating elements, and answering questions about content. Holo1.5 models are trained using a multi-stage pipeline with high-quality proprietary data, open-source datasets, synthetic data, and human-annotated samples, including large-scale supervised fine-tuning followed by online reinforcement learning (GRPO).
Key Capabilities
- State-of-the-Art UI Localization: Holo1.5-7B demonstrates leading performance in accurately identifying the positions of UI elements (buttons, text boxes, images) across various interfaces. It achieves an average 4.5% improvement in localization accuracy over previous models on benchmarks like Screenspot-V2, Screenspot-Pro, GroundUI-Web, Showdown, and WebClick.
- Superior Screen Content Understanding (QA): The model excels at comprehending the structure and functionality of user interfaces to answer questions effectively. It improves upon state-of-the-art models by 3.9% in GUI-focused question answering tasks on benchmarks such as ScreenQA Short, ScreenQA Complex, VisualWebBench, and WebSRC.
- High-Resolution Processing: Natively supports high-resolution inputs up to 3840 x 2160 pixels, enabling accurate interpretation and action performance on large, complex screens.
Good For
- Developing advanced Computer Use (CU) agents for automating digital tasks.
- Applications requiring precise user interface (UI) localization and interaction.
- Building systems that need to understand and reason about screen content and answer UI-based questions.
- Creating productivity tools that delegate complex digital tasks safely and efficiently.