PicoGPT Profile Banner
Pico AI Server and Pico AI Studio Profile
Pico AI Server and Pico AI Studio

@PicoGPT

Followers
566
Following
174
Media
35
Statuses
277

The easiest to install and use free open source and local AI on Apple Silicon. Fully Ollama and OpenAI-compatible. Built on MLX

San Francisco, CA
Joined December 2022
Don't wanna be here? Send us removal request.
@PicoGPT
Pico AI Server and Pico AI Studio
4 days
OpenAI GPT-oss is now available Mac App Store for free.
@OpenAI
OpenAI
12 days
We released two open-weight reasoning models—gpt-oss-120b and gpt-oss-20b—under an Apache 2.0 license. Developed with open-source community feedback, these models deliver meaningful advancements in both reasoning capabilities & safety.
1
1
7
@PicoGPT
Pico AI Server and Pico AI Studio
3 days
*works :).
0
0
0
@PicoGPT
Pico AI Server and Pico AI Studio
3 days
This work, but if you prefer using a GUI to run MLX models, check the Mac App Store link in the next tweet.
@ai_christianson
Andrew I. Christianson
4 days
Running mlx optimized models is easy 🙂
1
2
3
@PicoGPT
Pico AI Server and Pico AI Studio
4 days
Two updates in one day! Pico AI Server 1.2.11 with support for OpenAI GPT-oss was released earlier today. 1.2.12 was released just now with Jan v1 support for tool calling.
1
1
8
@PicoGPT
Pico AI Server and Pico AI Studio
4 days
Just submitted Pico AI Server 1.2.12 with support for Jan v1 to the App Store. Expect it to be available for free in the next few days.
1
0
1
@PicoGPT
Pico AI Server and Pico AI Studio
4 days
Big thank you to @JohnMai_Dev for bringing GPT-oss to MLX-Swift!.
0
0
2
@PicoGPT
Pico AI Server and Pico AI Studio
4 days
OpenAI updated the response format with a format they call Harmony. I'm still figuring out how to best support Harmony in Pico AI Server. You'll see some tags like "<|channel|>analysis. ". You can ignore those tags for now.
Tweet media one
1
0
1
@PicoGPT
Pico AI Server and Pico AI Studio
4 days
Download Pico Private AI Server here and run GPT-oss, DeepSeek, Qwen, Llama, Gemma 3, and more on your own local network. 100% private.
Tweet card summary image
apps.apple.com
‎Private LLM and VLM in 60 seconds. Turn any Apple-silicon Mac into a zero-setup, enterprise-ready AI server. One download gives your team a web chat interface, and 300 + free state-of-the-art...
1
0
1
@PicoGPT
Pico AI Server and Pico AI Studio
5 days
RT @alandao_ai: Jan-v1: a 4B model optimized for search use cases 😍. This model is specifically designed for search-related tasks and has b….
0
2
0
@PicoGPT
Pico AI Server and Pico AI Studio
5 days
Coming to Pico AI server soon!.
@jandotai
👋 Jan
5 days
Introducing Jan-v1: 4B model for web search, an open-source alternative to Perplexity Pro. In our evals, Jan v1 delivers 91% SimpleQA accuracy, slightly outperforming Perplexity Pro while running fully locally. Use cases:.- Web search.- Deep Research. Built on the new version
1
1
6
@PicoGPT
Pico AI Server and Pico AI Studio
7 days
Pico AI Server 1.2.10 is available in the Mac App Store. New in this version: Qwen 3 4B 2507 Instruct and Qwen 3 4B 2507 Thinking.
0
1
7
@PicoGPT
Pico AI Server and Pico AI Studio
7 days
Coming soon to Pico AI Server.
@JohnMai_Dev
John Mai
7 days
MLX Swift GPT OSS~
Tweet media one
0
1
5
@PicoGPT
Pico AI Server and Pico AI Studio
10 days
1.2.9 is available now in the App Store. This version fixes a Qwen 3 Coder Flash issue.
1
1
3
@PicoGPT
Pico AI Server and Pico AI Studio
12 days
Once that's done, I will need to submit the new version of Pico AI Server to the App Store. Apple's review usually takes 1-2 days (that's a bit slower than usual, probably due to vacation). Expect GPT-OSS to be available on Pico in the next few days.
0
0
0
@PicoGPT
Pico AI Server and Pico AI Studio
12 days
. merged into the Apple repo later today. Once that's done, the community will need to port the Python code to Swift. I don't see any PRs yet but expect people have already started working on that. Hopefully we'll see the port today or tomorrow.
1
0
0
@PicoGPT
Pico AI Server and Pico AI Studio
12 days
Pico AI Server is build on Apple's AI library MLX-Swift. There currently is no MLX implementation of GPT-OSS. The kind people at @lmstudio just opened a pull request for MLX Python. There are some tiny fixes needed but I assume it will be.
Tweet card summary image
github.com
Contribution by LM Studio Team. Initial implementation, more details to follow. Please note numerical precision considerations vs. the reference implementation.
1
0
0