New Llama.cpp Update Lets More AMD Laptops Run AI Locally

Llama.cpp is a free tool that lets ordinary computers run large language models — the technology behind chatbots like ChatGPT — entirely on your own hardware. No internet connection required. Your data stays private because nothing leaves your machine. It's popular with people who want AI without paying monthly fees or sharing personal files.

This latest update, called b10842, adds support for "gfx90c," a graphics processor built into many modern AMD chips, especially in budget and mid-range laptops. Previously, owners of these machines often had to use the slower CPU for AI tasks. Now, their computer's GPU can pitch in, making responses noticeably faster. Think of it as opening a previously locked door for a huge group of users.

For everyday people, this means an AI assistant that runs like a normal app — fast, offline, and free. Students can use it for homework help without Wi-Fi. Small businesses can process customer data without uploading it to a third-party server. Even hobbyists can build smart home tools that respond to voice commands without a cloud subscription.

The catch: you still need a model file, some of which are several gigabytes, and enough RAM to run it smoothly. This update doesn't make a weak computer instantly powerful, but it removes a major barrier. As local AI improves, more of us will enjoy the privacy and cost benefits of keeping AI at home.