Following last month’s Lemonade SDK 10.0 release that finally makes AMD Ryzen AI NPUs under Linux useful for running large language models (LLMs) where as before the Linux build could only target GPUs, released on Monday was Lemonade 10.1 with more enhancements to this local LLM solution.
Lemonade 10.1 is now available for Windows and Linux users to assist in running local AI apps primarily with AMD Ryzen and AMD Ryzen AI hardware while also being able to fallback for modern x86_64 CPUs.
With Lemonade 10.1 its new command line interface is described as a “much nicer way” of interacting with the service. The lemonade-server CLI now just lemonade is also “massively simplified”. The lemonade launch sub-command is one of the improved CLIs with Lemonade 10.1 with now making it easier to select a model from a list rather than having to type out full model names, async model loading is restored, support for API keys with “–api-key”, and more.
Lemonade 10.1 also allows for “supercharged” coding with improved lemonade launch codex|claude.
Lemonade 10.1 also now supports super resolution upscaling for generated images. Additionally, Lemonade 10,1 now supports Gemma 4 models on GPUs.
Downloads and more details on today’s Lemonade 10.1 release via GitHub. In addition to the open-source code there is also a Linux AppImage build as well as an Ubuntu PPA and Fedora RPM too.
