← all releases
PrismML·Jul 14, 2026·rolling·open source
Bonsai 27B
Bonsai 27B is a multimodal, 27-billion-parameter language model optimized for extreme on-device efficiency using 1-bit and ternary weight compression. It enables complex agentic workflows, multi-step reasoning, and tool use on consumer hardware like smartphones and laptops without requiring cloud connectivity.
LLMOn-device AIModel CompressionMultimodal
overview
Bonsai 27B is a multimodal, 27-billion-parameter language model optimized for extreme on-device efficiency using 1-bit and ternary weight compression. It enables complex agentic workflows, multi-step reasoning, and tool use on consumer hardware like smartphones and laptops without requiring cloud connectivity.
key features
- 011-bit and Ternary weight compression achieving a 3.9GB-5.9GB footprint
- 02Multimodal capabilities supporting vision tasks like screenshots, documents, and camera input
- 03Full 262K-token context window
- 04Optimized for on-device agentic loops and autonomous tool-calling
- 05Supports speculative-decoding for accelerated token generation
- 06End-to-end local execution via Apple MLX and NVIDIA CUDA
related productsbrowse all →
Protenixopen sourceProtenix is an open-source, trainable PyTorch implementation of AlphaFold 3, designed for high-accuracy biomolecular structure prediction. It aims to advance accessible and extensible research tools for the computational biology community. ([github.com](https://github.com/bytedance/Protenix?utm_source=openai))Z-Imageopen sourceZ-Image is an efficient image generation foundation model developed by Tongyi-MAI, designed to produce high-quality, diverse, and stylistically versatile images. It serves as a robust base for creators, researchers, and developers seeking advanced image generation capabilities.