Long-horizon agentic work
Hy4 Preview is positioned for long-horizon agentic tasks.
Productivity / product dossier
An open 770B MoE model for long-horizon coding and document analysis.
Product brief
Hy4 Preview is a 770B MoE model (49B active, 1M context) from Tencent, built for long-horizon agentic tasks. It autonomously handles coding, game dev, and complex document analysis, running its own tests and fixing bugs before delivery.
Why we selected it
A technically notable open 770B MoE model with 1M context, positioned for autonomous coding, testing, bug fixing, game development, and document analysis.
Product preview saved with our daily selection
Capability scan
Only capabilities supported by the product information we collected are listed here.
Hy4 Preview is positioned for long-horizon agentic tasks.
It is described as autonomously handling coding, running its own tests, and fixing bugs before delivery.
The model is positioned for game development work.
The model is positioned for analyzing complex documents.
Best-fit use cases
FAQ
Hy4 Preview is a Tencent model positioned for long-horizon agentic tasks.
The description identifies it as a 770B mixture-of-experts model with 49B active parameters.
The stated context length is 1M.
The description lists coding, game development, complex document analysis, running tests, and fixing bugs before delivery.
Yes. The supplied editorial selection reason characterizes it as an open 770B MoE model.