H Company's new Holo2 model takes the lead in UI Localization

Hugging Face Blog · 5 months ago

H Company released Holo2-235B-A22B Preview, a UI localization model that achieves 78.5% accuracy on the Screenspot-Pro benchmark and 79.0% on OSWorld G. The model reaches 70.6% accuracy in a single step but improves to 78.5% within three iterative refinement steps, representing a 10-20% relative gain across all Holo2 model sizes. The model is available on Hugging Face as a research release focused on identifying UI elements in high-resolution 4K interfaces.