Llama.cpp on a jailbroken PS5: run any GGUF model natively on the console

A developer has created PS5LM, a tool that enables jailbroken PlayStation 5 consoles to run large language models locally using llama.cpp on the console's GPU, serving them as an API to devices across a home network. The system supports GGUF model formats, offers OpenAI and Anthropic-compatible APIs, can run decision models, and displays real-time performance metrics on a TV, requiring no PC, cloud, or external accounts.