AMD’s latest ROCm update adds support for new Ryzen APUs and enhances local AI features.
This consistent progress in ROCm improvements demonstrates how local AI deployment is rapidly gaining power.
These software advances stem from AMD’s focused efforts over recent years to improve the ROCm stack, particularly for Edge AI. At CES2026, for example, AMD introduced ROCm 7.2.7, which supports the new Ryzen AI 400 Gorgon Point APUs. Building on this, the company has also improved local model performance, which we’ll cover soon.
In addition to these hardware advances, AMD has prioritized seamless ROCm integration with tools like Comfy UI, an image generation tool that delivers a 5-fold performance increase in ROCm 7. The company has also expanded ROCm support for consumer products, effectively doubling Ryzen and Radeon compatibility over the past year. Together, these moves illustrate how AMD’s software strategy aligns with its consumer objectives.
Reflecting this increased user base, AMD has introduced smooth integration with the ONNX path for inference and training, targeting Windows AI users and OEMs. Additionally, ROCm is now compatible with PyTorch on Windows, and the Rock software package, an open-source platform from HIP and ROCm. Through these steps, Windows is becoming a key platform for ROCm as AMD drives local AI adoption into the mainstream.
These technical and strategic improvements now enable local AI inference on consumer hardware to nearly match the quality of cloud-based models. For instance, AMD compared open-source models like GPT running on Ryzen AI Max Plus APUs against cloud-based counterparts. According to AMD, results show that for benchmarks similar to GPQA Diamond and MMLU, local and cloud performances are largely comparable, highlighting how much edge AI has improved through ROCm and new hardware capabilities.
Source: AMD ROCm 7.2.2 Adds Support for Ryzen AI 400 CPUs & Unlocks Faster Local Inference Performance
