Centering AI research on efficiency. https://xitter.cf/t.co/88MQHGCeFD
United States
Joined March 2025
- Tweets408
- Following32
- Followers29.6K
- Likes628
Pinned Tweet
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Small, but mighty 🌳
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Getting a model onto a phone is one milestone. Making it useful enough to stay there is the next.
With Bonsai 2, we focused on the everyday stuff: following instructions, calling tools, and getting to the answer without rambling.
What would you want it to handle locally?
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Run Ternary Bonsai 2 locally in Atomic Chat 🪴
@PrismML shrank Qwen3.5 7x so the 9B goes from 17.9 GB to just 2.42 GB and it keeps up or outperforms the original model on benchmarks
Run AI models locally -> atomic.chat
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Bonsai 2 9B is LIVE on Darkbloom from day one.
@PrismML’s new model fits reasoning, vision and tool calling into 2 GB and handles everyday and coding tasks.
It runs on a network of Macs owned by everyday people.
The first 100 people get 250M free tokens at darkbloom.ai.
Want it faster? Bonsai 2 9B is also an open challenge at mlx.fast and cuda.fast on @YukonResearch.
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Had the pleasure of testing the new Bonsai 2 2/4/9B models recently and here are my results
9B is truly exceptional for its size, but 4B is where it’s at. Also no more multiple thinking modes, just on/off so no need to worry about excessive capping we saw with original.
2B is bit of mixed bag, but what it does well, it does it consistently while being fraction of size as 27B. Can run many of these even on a small GPU, so highly recommend testing it out to see how it fits your use cases
PrismML retweeted
The new Bonsai 2 2B model reaches up to 120 tk/s on iPhone 18 Pro
It’s better than Qwen 3.5 2B (4-bit) and runs twice as fast while using almost half the memory
PrismML retweeted
Congratulations @PrismML on releasing Bonsai 2 2B, 4B and 9B 👏👏
If you are building AI products with tight memory, latency, or power requirements, reach out at [email protected]
@BabakHassibi @HessianFree @SahinLale
cc @khoslaventures @vkhosla
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
We’re also launching an optimization competition with @darkbloomai to make Bonsai 2 9B faster. Check it out on @YukonResearch!
PrismML retweeted
@PrismML continues to deliver. Love the work these guys do to bring quality AI to everyone on the most common devices.
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Do you have 2 to 4 GB of vram?
Great! Now you can run a lil guy
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Today we're excited to extend the Bonsai 2 family of models with ternary 2B, 4B, and 9B models based on Qwen 3.5. These may not solve IMO problems or the Riemann hypothesis but are incredible for high volume and high speed tasks on everyday devices.
The Bonsai 2 4B model runs 32 concurrent streams on my Macbook Pro (M5 Pro) to process 250 emails and parse them into JSONs with summaries, action items, calendar events, and priority scoring. The peak memory use is less than 4GB and in total generate over 700 tokens per second.
PrismML retweeted
"Solving the Riemann hypothesis can wait." 😂
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
🚨 JUST LAUNCHED: Bonsai 2 from @PrismML.
Open-weight ternary models in 2B, 4B and 9B, small enough to run on a phone. The 2B hits 90 tokens/sec on iPhone 17 Pro Max, with reasoning, vision and tool calling all on-device.
You can try Bonsai 2 9B today on @DarkbloomAI, where the first 100 people get 250M free tokens.
And we’re making it faster on @YukonResearch.
Bonsai 2 9B runs at 170+ tokens/sec at launch.
There are two tracks, Apple Silicon on mlx.fast and NVIDIA DGX Spark on cuda.fast.
To join, pick a track, point your agent at it and submit a faster run.
Every new best is open for anyone to build on.
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
New challenge on Yukon: Make Bonsai 2 9B faster.
@PrismML’s new Bonsai 2 9B packs reasoning, vision and tool calling into just 2 GB. It runs on an iPhone at up to 22 tokens per second.
On a Mac, it starts at 170+. We think that’s the floor.
Two tracks:
- Apple Silicon on mlx.fast
- NVIDIA DGX Spark on cuda.fast
Optimize it, climb the leaderboard, and every new best is open for anyone to build on.
Bring your agent.
Bonsai 2 9B, 4B, and 2B are available under Apache 2.0. We look forward to seeing what you build.
Blog: prismml.com/news/bonsai-2-fa…
Models: huggingface.co/collections/p…
Darkbloom API: console.darkbloom.dev/models
Demo: prismml-eng.github.io/Talk/
Try on iPhone: apps.apple.com/us/app/locall…
Instructions and Examples: github.com/PrismML-Eng/Bonsa…
Docs: docs.prismml.com
Whitepaper: github.com/PrismML-Eng/Bonsa…
Discord: discord.gg/prismml
Join us: prismml.com/careers
Alongside the release, we're launching an open challenge on @YukonResearch to make Bonsai 2 9B even faster. You can also try it today on @DarkbloomAI.
PrismML retweeted
Maybe the toaster thing will come true.
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Today has been one of the biggest days for open-source AI this year... and it keeps getting better.
@PrismML just released more Ternary Bonsai 2 models on Hugging Face: now in 2B, 4B, 9B, and 27B. The 2B reaches up to 500 tok/s on my M4 Max ENTIRELY in the browser with custom WebGPU kernels... that's insane! 🤯
At ~1.8 bits per weight, the language-model weights are just 430MB for 2B and 950MB for 4B (plus ~365MB for the 8-bit vision weights). The future of local agentic AI is going to be wild.
Try out the demo yourself 👇
PrismML retweeted
Today we are releasing three new members of the Ternary Bonsai 2 family. Built from Qwen 3.5, they come in 2B, 4B, and 9B varieties and are more than 9x smaller than their full precision counterparts.
An advantage of having open sourced our previous models is the feedback that we have received (both good and bad). The community has stretched our models in every which way, exposing flaws, some of which, would have taken months of internal evaluation to reveal.
We have done our best efforts to address these. In particular, we have focused on what makes small models useful as local assistants. We have improved instruction-following and tool-calling on mobile devices, while attempting to avoid repetitive loops and providing useful answers without unnecessary output.
All three models are under Apache 2.0.
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
Excited about releasing these small VLMs. We’ve been hard at work trying to bring more intelligence to your devices. We were thoughtful in building these models. We accounted for typical use cases for each model size. Looking forward to seeing what great apps you all build!
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.
PrismML retweeted
👀 oh lawd have mercy
Today, we’re releasing Bonsai 2 2B, 4B, and 9B.
Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices:
• 2B: 0.43GB
• 4B: 0.95GB
• 9B: 2.0GB
For this release, we focused on what makes small models useful as local assistants: following instructions, avoiding repetitive loops, calling tools on mobile devices, and giving useful answers without unnecessary output. The 9B model uses roughly 40% fewer tokens on standard instruction-following and everyday math tasks. Solving the Riemann hypothesis can wait.
These models are successors to our first-generation Bonsai 1.7B, 4B, and 8B, and provide a foundation for bringing practical agentic functionality to local applications on personal computers, phones, and wearables.
The Bonsai 2 family now includes four sizes spanning 2B to 27B parameters. Developers can choose the balance of capability, memory use, and responsiveness that fits their application and hardware.
Available today under Apache 2.0.