Posts

@BennettBuhner and this is only the 20b model?? woah

@uwukko 😭😭i like how they said it runs on “most desktops/laptops” or whatever they said

@srt54558 Ik but Sam mentioned it running on a phone and this seems to have 3b active params or smth

@rodriveiga100 https://x.com/jglypt/status/1952803418570985937 https://x.com/jglypt/status/1952803418570985937

@ollidotcom https://x.com/jglypt/status/1952803418570985937 https://x.com/jglypt/status/1952803418570985937

Didn’t work… the app developer will prob have to update it to support it (based on previous models when they first dropped)

on Enclave btw

Well… downloading gpt-oss 20b on my iPhone 16.. lets see how (slowly) this’ll run (walk) 🤔

<img src=“https://jackk.micro.blog/uploads/2025/0b845ded93.png" width=“276” height=“600” alt=“A user interface displays a model manager with options such as “Gemma 3 - 4B,” “Gemma 3n E2B - 2B,” and “LiquidAI LFM2 - 350M,” each showing parameters and creators.">

@markgurman I read that as “5s” like an iPhone 5s :p

@sama I feel like the 20b OSS model won’t really work well on a phone… 😅