Pasha

Engineering @PrismML

[Pasha @pashakho]

@anemll Yeah, depends on the task for speed-up. Tradeoffs for dspark are a bit different for 1-bit models as well.

[Pasha @pashakho]

@pupposandro thanks :)

[Pasha @pashakho]

Bonsai-27B ships with a little friend: a dspark drafter checkpoint 🌱⚡

Same model, same prompt, one L40S GPU: 72 → 135 tok/s

Flip it on and go zoom 🏎️💨

[Pasha @pashakho]

Bonsai in the cloud 🌳☁️🖥️

[AK @_akhaliq]

Bonsai 27B now available in claude code via hf claude

[Pasha @pashakho]

@Tono_Ken3 Ternary Bonsai should work on LM Studio, slightly larger though. Working on supporting the 1-bit soon.

https://t.co/k2cVZ2aSDg

[Together AI @togethercompute]

Bonsai 27B is a big step for local AI: 27B-class multimodal capability in a phone-class footprint.

Congrats to @PrismML on the launch.

Try Ternary Bonsai 27B on Together AI: https://t.co/yeyvZcaATO

[Pasha @pashakho]

Try it yourself, one-command setup 🛠️👇
https://t.co/HzjKwUQV56

[Adrien Grondin @adrgrondin]

The new PrismML’s Bonsai 27B model running on iPhone 17 Pro

Bonsai 27B is the first 27B-class model to run on a phone and it uses less than 5GB of memory

Pushing the limits on what can run a phone

[Pasha @pashakho]

Ternary Bonsai 27B on a Mac: open chart, summon tools, investigate the stocks 🌱💻📈

SNDK and MU go in. Plots and analysis come out.

Demo only 🧪 Not financial advice.

[Pasha @pashakho]

@xenovacom @PrismML Awesome thanks, great demos as usual 🔥

[Pasha @pashakho]

Bonsai 27B go fast 🚀🌳

[NVIDIA RTX Spark @NVIDIARTXSpark]

Congratulations to @PrismML on the launch of Bonsai 27B family of models!

Try them out today on NVIDIA RTX GPUs & DGX Spark.

[Pasha @pashakho]

Hugging Face collection: https://t.co/kD4Gf13H6h

Github Demos:
https://t.co/HzjKwUQV56

[Pasha @pashakho]

Bonsai 27B 🌳🌳🌳

[Pasha @pashakho]

🌳🌳🌳🔥🔥🔥

[Xenova @xenovacom]

Bonsai 27B just changed the local LLM game forever.

1-bit quantization shrinks it from 54GB to just 3.8GB (-93%), while retaining 90% of its intelligence. That's insane.

With custom WebGPU kernels written by Fable 5 and GPT 5.6 Sol, the model now runs locally in your browser!

[Pasha @pashakho]

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.
https://t.co/3QFvwYdGcj

[PrismML @PrismML]

Today, we’re announcing Bonsai 27B: the first 27B-class model to run on a phone.

Bonsai 27B is the new multimodal flagship of the Bonsai family. Based on Qwen3.6 27B, it brings a new capability tier to local AI: multi-step reasoning, structured tool use, long-context workflows, and coherent agentic loops.

Until now, models in this class have been impractical to deploy locally. A 27B model occupies roughly 54 GB in 16-bit precision, and even a strong 4-bit build is around 18GB - too large for a phone and for most laptops.
Bonsai 27B changes that.

It comes in two variants:
• Ternary Bonsai 27B: 5.9 GB, 1.71 effective bits per weight, optimized for laptop-class quality.
• 1-bit Bonsai 27B: 3.9 GB, 1.125 effective bits per weight, optimized for phone-class footprint.

Everything is open-sourced today under the Apache 2.0 license.

[Locally AI - Local AI Chat @LocallyAIApp]

Meet 1-bit Bonsai 27B, the latest model from @PrismML and the first 27B-class model to run on a phone.

This model, available on the iPhone 17 Pro, iPhone Air, and select iPads, pushes the limits of intelligence density. Powered by MLX.

Available now.

[Years Progress @YearsProgress]

2026 is 50% complete.

[Psyho @FakePsyho]

If you're sad that Fable 5 is no longer accessible, I'm happy to report that Fable 1, Fable 2 & Fable 3 are still available via Game Pass

[lcamtuf @lcamtuf]

Did you know? The latest edition of the C standard introduces left-worm and right-worm operators which can be used in place of the cumbersome x + 1 and x - 1 notation: https://t.co/k5WGvCT3Iv

[Awni Hannun @awnihannun]

The video from @angeloskath on local agentic AI with MLX is excellent. I also hear it's one of the most viewed videos in WWDC history 👏

Goes through the basics of agentic AI and how to set it all up to run locally in a very approachable and simple way.

The demos are excellent and it's kind of wild that they just work now. None of this was possible or practical < 1 year ago before M5 and the recent quality bump in open weights models. And it's not done improving.