PrismML Releases Bonsai Image 4B
Bringing high-quality image generation to local devices
Available in both 1-bit and ternary variants, Bonsai Image 4B reduces the footprint of a modern 4B-class diffusion transformer by up to 8.3x while preserving strong visual quality. This reduction makes Bonsai Image 4B, the first image model in its parameter class to run directly on the iPhone. The models are built for local inference across iPhone, Apple Silicon Macs, CUDA GPUs, and local or small-scale serving environments.
"Local image generation is the next major milestone for creative AI," said
About Bonsai Image 4B
Bonsai Image 4B is available in two variants:
- 1-bit Bonsai Image 4B uses binary {−1, +1} transformer weights with group-wise FP16 scaling. It targets maximum compression and is designed for deployments where memory pressure and model footprint are the primary constraints.
- Ternary Bonsai Image 4B uses ternary {−1, 0, +1} transformer weights with group-wise FP16 scaling. The additional zero state gives the model more representational flexibility, improving visual quality and prompt fidelity while remaining extremely compact.
The 1-bit variant compresses the diffusion transformer to 0.93 GB, an 8.3x reduction from the full-precision model. The ternary variant compresses it to 1.21 GB, a 6.4x reduction from the full-precision transformer. The compressed variants retain up to 95% of the image-generation quality of the full-precision model.
On iPhone 17 Pro Max, Bonsai Image 4B generates a 512x512 image in about 9.4 seconds. On Mac M4 Pro, the same resolution takes about 6 seconds. On Mac M4 Pro, generations from Bonsai Image 4B are up to 5.6x faster than the stock full-precision pipeline.
Making local image generation practical
Cloud image generation will continue to be the right choice for many products. But cloud-only generation imposes product constraints: every prompt is a remote request, every iteration has marginal serving cost, and every interaction adds round-trip latency.
That matters because image generation is naturally iterative. Users revise prompts, compare outputs, generate variations, discard failures, and try again. When every attempt depends on a server-side call, teams often have to ration generations, introduce credits or rate limits, or design the experience around fewer attempts.
Local inference changes that. Once the model fits on the device, generation can sit directly inside the product experience. It becomes cheaper to run, faster to iterate with, and easier to use in environments where prompts and generated assets should remain private.
Pricing and availability
Both 1-bit and Ternary Bonsai Image 4B will be released with open weights and code under the Apache 2.0 license.
With this launch, PrismML is also introducing
Resources
- Launch Blog
- Download Bonsai Image models
- Bonsai Image demo
Download Bonsai Studio for iPhone
- Whitepaper
- GitHub
About PrismML
PrismML is a
All registered trademarks and product identifiers belong to their respective corporate entities. Any other trademarks or product names referenced here are also owned exclusively by their relevant companies.
Media Contact
PrismML
[email protected]
831.888.9011
View original content:https://www.prnewswire.com/news-releases/prismml-releases-bonsai-image-4b-302782354.html
SOURCE PrismML
Serious News for Serious Traders! Try StreetInsider.com Premium Free!
You May Also Be Interested In
- Tata Elxsi and Sky Mark a Major Milestone in AI-led Autonomous Network Transformation with NEURON
- Premier Development & Investment, Inc. Now Confirms Certain Highly Material Transactions to Close Imminently and Extends its Extreme “Cautionary” to All Parties
- Vision Marine Technologies Receives USPTO Notice of Allowance for Electric-Vessel Powertrain Authentication Technology
Create E-mail Alert Related Categories
PRNewswire, Press ReleasesRelated Entities
Cerberus CapitalSign up for StreetInsider Free!
Receive full access to all new and archived articles, unlimited portfolio tracking, e-mail alerts, custom newswires and RSS feeds - and more!



Tweet
Share