One click install local AI server. Discord: https://xitter.cf/t.co/bbFV10wiKl
San Francisco, CA
Joined December 2022
- Tweets342
- Following596
- Followers814
- Likes208
Pinned Tweet
Apple Silicon + Gemma 4 fans: this is for you.
Pico AI Server now supports continuous batching with MLX-Swift.
43 tok/s on 1 stream.
26 tok/s per stream on 2 concurrent streams.
That’s 52 tok/s total. a 21% throughput gain on a six-year-old MacBook Pro M1 Max!
👇
Mark your calendars: the AiOS Meetup is returning next Tuesday in San Francisco!
@Hailuo_AI will demonstrate how to run the new H3 video model locally and free on a Mac
@JoannisOrlandos of @wendylabsinc will show how to deploy AI on robots and embedded systems using Swift
The event will be hosted by @RayFernando1337
Big thank you to @waynesutton of @convex for sponsoring the event
Pico AI Server and Pico AI Studio retweeted
These models are super impressive. Tiny yet accurate and surprisingly good at tool calling.
I'm preparing an update to @PicoGPT that will use Ternary Bonsai 27B as default small model. If you can't wait, install it manually. Settings -> Models -> + -> Add Models from Web URL -> huggingface.co/prism-ml/Tern…
We grew a very big model in a very tiny pot. 🌱
Meet Bonsai 27B; a 27B-class multimodal model that fits on a phone.
🪴 3.9 GB at 1-bit
💻 5.9 GB in ternary
🧠 Reasoning, vision, tool use
🔓 Apache 2.0
Small footprint. Full-grown intelligence.
prismml.com/news/bonsai-27b
Pico AI Server and Pico AI Studio retweeted
AI apps get more useful when models can call tools, but local models often pick the wrong one(s).
Last week @LiquidAI shipped a tiny new retrieval model that picks the right tools before the query hits the LLM.
It's now on MLX: running fully local on Apple Silicon. 🧵
Congrats @MiniMax_AI on the release! No word yet on the number of parameters, but Pico will support M3 when the weights are available on HuggingFace.
Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities
- Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas
- MiniMax Sparse Attention scales context to 1M
- Natively Multimodal from Step Zero
API: platform.minimax.io
Token Plan: platform.minimax.io/subscrib…
🚀New! MiniMax Code: code.minimax.io
Weights & Tech Report in ~10 Days
Pico AI Server and Pico AI Studio retweeted
Huge shoutout to @PrismML
This shouldn’t be possible: a tiny model punching way above its weight.
The largest version is just 1.14 GB, which means it’s small enough for a phone.
Fast on a phone (spoiler: Pico for iOS is coming soon!). Insanely fast on a MacBook Pro M1 Max.
Pico Local AI Server 1.4.21 is now available on the Mac App Store. This release adds support for Ternary Bonsai, a lightning-fast model that outperforms many much larger models
Today we’re announcing Ternary Bonsai: Top intelligence at 1.58 bits
Using ternary weights {-1, 0, +1}, we built a family of models that are 9x smaller than their 16-bit counterparts while outperforming most models in their respective parameter classes on standard benchmarks.
We’re open-sourcing the models under the Apache 2.0 license in three sizes: 8B (1.75 GB), 4B (0.86 GB), and 1.7B (0.37 GB).
Download or update Pico AI Server today:
apps.apple.com/us/app/pico-l…
Pico AI Server and Pico AI Studio retweeted
おー、ためす
Pico AI Server and Pico AI Studio retweeted
Excited to share that @awnihannun, co-creator of MLX, is joining AiOS Meetup 2026 for a special talk in Cupertino on June 11
Come hang with @RayFernando1337, @rudrank and folks building at the intersection of AI and Apple platforms. RSVP here:
luma.com/ju8j7vtp
Pico AI Server and Pico AI Studio retweeted
If you haven’t already, please rate the app on the App Store. It helps more people discover it. We’re at 4.6 in the U.S. and 4.5 worldwide, help us raise that
apps.apple.com/us/app/local-…
Welcome to all our new followers! Need support or have a feature request? Join our Discord: discord.com/invite/Nrf5y8Uax…
If you haven’t already, please rate the app on the App Store. It helps more people discover it. We’re at 4.6 in the U.S. and 4.5 worldwide, help us raise that
apps.apple.com/us/app/local-…
Pico AI Server and Pico AI Studio retweeted
Pico AI Server another one to be added to my engine benchmarks! Be ready for some feedback @ronaldmannak 💪
Pico AI Server and Pico AI Studio retweeted
Apple Silicon + Gemma 4 fans: this is for you.
Pico AI Server now supports continuous batching with MLX-Swift.
43 tok/s on 1 stream.
26 tok/s per stream on 2 concurrent streams.
That’s 52 tok/s total. a 21% throughput gain on a six-year-old MacBook Pro M1 Max!
Pico update: @MiniMax__AI M2.5 is now available on Apple Silicon Macs with 192GB or more unified memory
