The model accepts visual and textual context and can reason about charts, screenshots and other images alongside written instructions. Compatible applications connect the model to command-line or graphical tools so it can plan and carry out multi-step work.
Developers can use Hy4 through Tencent Cloud TokenHub to build image-aware assistants or integrate it with existing coding and office tools. Tencent describes the final hosted model separately from the downloadable Hy4 Preview checkpoint.
