Open Models · first seen 13 October 2026, 22:21 UTC
Jailbroken PS5 runs Qwen 27B at ~20 tok/s
An enthusiast ran a 27 billion parameter Qwen model on a jailbroken PlayStation 5's GPU at about 20 tokens per second, demonstrating local LLM inference on console hardware.
Open
- Params
- 27B