gg2
Llama.cpp runs GGUF models on jailbroken PS5
promptzone.com

Llama.cpp runs GGUF models on jailbroken PS5

The gist

A jailbroken PS5 ran a 27-billion-parameter Qwen model locally at about 20 tokens per second via PS5LM Plus.
The port brings llama.cpp's full range of GGUF models to console hardware.

How it unfolded

  1. Oct 14, 2026@Pirat_Nation posted on X that a jailbroken PS5 was running Qwen 3.8 27B locally at around 20 tokens per second.
  2. Oct 15, 2026ProPakistani reported the 27-billion-parameter model running on the console's GPU via PS5LM Plus, a fork of Mert Cobanov's PS5LM project.
  3. Oct 16, 2026PS5LM v0.9.0 was released, running llama.cpp on the PS5 GPU as its own app with a redesigned TV interface.

Sources

Read the full story in the app gg2 — free on the App Store