<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/DzP294tP6" -->

---
title: Jailbroken PS4 runs 9-billion-parameter AI model at up...
description: An open-source project called Garlic LLM runs a 9-billion-parameter AI model directly on a jailbroken PS4's 2013-era GPU, offline. Built by developer Leonardo...
canonical: https://daily.dev/posts/dzp294tp6
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Jailbroken PS4 runs 9-billion-parameter AI model at up to 43 tokens per second — open-source Garlic LLM bypasses drivers to talk directly to the 2013-era GPU | daily.dev
og:description: An open-source project called Garlic LLM runs a 9-billion-parameter AI model directly on a jailbroken PS4's 2013-era GPU, offline. Built by developer Leonardo...
og:url: https://daily.dev/posts/dzp294tp6
og:image: https://api.daily.dev/og/posts/DzP294tP6.png
og:image:alt: Post cover image
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Jailbroken PS4 runs 9-billion-parameter AI model at up to 43 tokens per second — open-source Garlic LLM bypasses drivers to talk directly to the 2013-era GPU

**[Jacob B. Bonde](https://daily.dev/sources/5cqvizkr5tdfdotvdjulg)** · [@byteoutlaw](https://daily.dev/byteoutlaw) · 38 upvotes · 3 comments

## Summary

An open-source project called Garlic LLM runs a 9-billion-parameter AI model directly on a jailbroken PS4's 2013-era GPU, offline. Built by developer Leonardo Alegría, it runs Alibaba's Qwen3.5-9B at a claimed 31 to 39 tokens per second in text chat and up to 43 for code, using 4-bit and 3-bit quantization to fit the model into about 4.5GiB. It talks to the GPU directly, building PM4 command buffers itself and using GCN compute kernels written in OpenCL C, and it runs on a PS4 with GoldHEN homebrew, tested on firmware 11.52. It also supports voice chat via Chatterbox and dictation via Nvidia's Parakeet TDT 0.6B v3, and is a free 5.66GB download on Hugging Face.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.tomshardware.com/video-games/playstation/jailbroken-ps4-runs-9-billion-parameter-ai-model-at-up-to-43-tokens-per-second-open-source-garlic-llm-bypasses-drivers-to-talk-directly-to-the-2013-era-gpu>

## Community discussion

Top comments from developers on daily.dev.

**@andreasenni** · 1 upvotes

> please AI don't start ruining consoles too

**@kizomanizo** · 0 upvotes

> I know what I'm doing with my dusty, shelved 4 slim.

**@randy** · 0 upvotes

> Ha. Mine is still playing games.. But I guess that's what I get for not getting a PS5 when they were _reasonably_ priced. 😞

## Similar posts on daily.dev

- [Llama.cpp on a jailbroken PS5: run any GGUF model natively on the console](https://daily.dev/posts/llama-cpp-on-a-jailbroken-ps5-run-any-gguf-model-natively-on-the-console-ehegpu47t) · Hacker News · 1 upvotes · 0 comments
- [I replaced ChatGPT with a local model on my gaming PC, and it's beating the cloud where I didn't expect](https://daily.dev/posts/i-replaced-chatgpt-with-a-local-model-on-my-gaming-pc-and-it-s-beating-the-cloud-where-i-didn-t-exp-boruivfcs) · XDA Developers · 0 upvotes · 0 comments
- [AI On Your Gaming PC](https://daily.dev/posts/ai-on-your-gaming-pc-a7iadecz7) · Hackaday · 2 upvotes · 2 comments
- [I Was Waiting for an RTX 3090 to Run Local AI. My Old RTX 2070 Changed My Plans.](https://daily.dev/posts/i-was-waiting-for-an-rtx-3090-to-run-local-ai-my-old-rtx-2070-changed-my-plans--ndwfrxh86) · Medium · 0 upvotes · 0 comments

---

Tags: [#data-science](https://daily.dev/tags/data-science), [#ai-inference](https://daily.dev/tags/ai-inference), [#reverse-engineering](https://daily.dev/tags/reverse-engineering), [#opencl](https://daily.dev/tags/opencl)

[View this post on daily.dev](https://daily.dev/posts/dzp294tp6)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/dzp294tp6","headline":"Jailbroken PS4 runs 9-billion-parameter AI model at up to 43 tokens per second — open-source Garlic LLM bypasses drivers to talk directly to the 2013-era GPU","text":"Shared: Jailbroken PS4 runs 9-billion-parameter AI model at up to 43 tokens per second — open-source Garlic LLM bypasses drivers to talk directly to the 2013-era GPU","url":"https://daily.dev/posts/dzp294tp6","datePublished":"2026-11-16T13:42:17.248Z","dateModified":"2026-11-16T14:32:25.944Z","author":{"@type":"Person","name":"Jacob B. Bonde","url":"https://daily.dev/byteoutlaw","image":"https://media.daily.dev/image/upload/s--veTChHK7--/f_auto/v1733220070/avatars/avatar_5cQvIZKr5tDFDotVDjulg","description":"I am about to change this world with my fellow programmers! 😊❤️","interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"EndorseAction"},"userInteractionCount":27025}},"image":"https://media.daily.dev/image/upload/s--ZrL_HSsR--/f_auto/v1722860399/public/Placeholder%2006","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":38},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":3}],"sharedContent":{"@type":"WebPage","url":"https://api.daily.dev/r/q3ywTDkP3"},"comment":[{"@type":"Comment","text":"please AI don’t start ruining consoles too","datePublished":"2026-11-17T07:39:25.818Z","url":"https://daily.dev/posts/DzP294tP6#c-5241Qqn5b","author":{"@type":"Person","name":"Andrea Senni","url":"https://daily.dev/andreasenni","image":"https://lh3.googleusercontent.com/a/ACg8ocL9zDvgvAHJyMsmnVewPaOwW1fMuVIk0D1besTkxG8tnUmIoQ=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"I know what I’m doing with my dusty, shelved 4 slim.","datePublished":"2026-11-17T09:25:23.107Z","url":"https://daily.dev/posts/DzP294tP6#c-UfFsCrkfx","author":{"@type":"Person","name":"Kizito","url":"https://daily.dev/kizomanizo","image":"https://lh3.googleusercontent.com/a/AAcHTtf_9jV6bHsBs8P3XMkgVR571XI3Mth0K9aXVStyQc3Rb-8=s96-c"}},{"@type":"Comment","text":"Ha. Mine is still playing games… But I guess that’s what I get for not getting a PS5 when they were reasonably priced. 😞","datePublished":"2026-11-20T14:35:50.847Z","url":"https://daily.dev/posts/DzP294tP6#c-tOe2iKwc2","author":{"@type":"Person","name":"Randy","url":"https://daily.dev/randy","image":"https://media.daily.dev/image/upload/s--UjV4-KkB--/f_auto/v1708097210/avatars/avatar_HXYbbGcBO38Rfv7RrCBdA"}}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/5cqvizkr5tdfdotvdjulg","name":"Jacob B. Bonde"}}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Jacob B. Bonde","item":"https://daily.dev/sources/5cqvizkr5tdfdotvdjulg"},{"@type":"ListItem","position":3,"name":"Jailbroken PS4 runs 9-billion-parameter AI model at up to 43 tokens per second — open-source Garlic LLM bypasses drivers to talk directly to the 2013-era GPU"}]}
```

