H company has released Holo4, a new family of agentic models designed to handle software through any available interface. The lineup includes a 27B dense model and a 35B-A3B mixture-of-experts model, both available through the H Models API, alongside an updated Holotron4 Nano. The models are built for real business workflows rather than academic benchmarks alone, and were trained with supervised and reinforcement learning on a large set of environments and tasks, including those generated by the company's Agentic Task Factory.
Unlike most agentic models, which are typically trained for a single interface, Holo4 can click and type on a screen, write and run code, and call MCP or API tools. The same model works on desktops, the web, Android, code sandboxes, and business APIs, so users do not need to pick a different model per platform. The company says this reflects real work, where a single task often spans multiple interfaces.
On the OSWorld 2.0 desktop-control benchmark, Holo4 27B scores 61.7% and the 35B-A3B reaches 30.9%, compared with 81.8% for Opus 5.5. The models also perform competitively on the AutomationBench API-use benchmark. The company emphasizes that this comes at a fraction of the parameter count and cost of frontier closed models. Every trajectory behind the public benchmark scores is open-sourced, allowing anyone to replay the steps or download the data from Hugging Face.