Developers can now build and test AI software on the AMD Ryzen AI Halo Developer Platform using Microsoft’s Project Zenith. The Windows 11 setup comes with coding tools already in place and AMD guides for tasks such as running a language model or generating images on the computer itself.
Ryzen AI Halo is the first device to get Project Zenith since AMD and Microsoft announced it at IFA 2026. It is available on the new Ryzen AI Halo powered by the AMD Ryzen AI Max+ PRO 495 processor. Other devices with Ryzen AI Max processors will get Project Zenith soon.
Project Zenith puts Visual Studio Code on the Windows taskbar and includes Git, a tool developers use to track changes to their code. Python, Node.js, .NET, and PowerShell 7 are also ready to use. Developers who work with Linux tools can access them through the Windows Subsystem for Linux.
The setup includes command-line tools for GitHub, GitHub Copilot, and Azure. Microsoft also configures Windows 11 settings for development. Developers can change those settings and add the tools they prefer.
For AI projects, the AMD Ryzen AI Developer Center offers software configurations that AMD has tested. Its step-by-step guides, called AMD AI Playbooks, show how to run large language models using PyTorch and AMD ROCm software. One guide pairs Visual Studio Code with Qwen3-Coder, an AI model that can help write code locally.
Other guides cover generating images with ComfyUI, making a language model available to applications through LM Studio, and using a local model with n8n to automate tasks.
The hardware helps determine which models developers can run on the PC. Ryzen AI Halo has up to 192 GB of shared memory, of which up to 160 GB can go to graphics. It can move data through memory at up to 273 GB per second, according to AMD.
AMD said the system can run models with up to 300 billion parameters. Parameters are the values a model learns during training. The available memory also gives developers room to work with longer passages of text or use several models in the same project. The system uses the x86 architecture found in many Windows PCs, so developers can keep using familiar applications while testing AI software on the same machine.