PrismML launches Bonsai 2 27B: 9x size compression while still retaining 98.2% of Qwen3.8 27b’s capabilities.
AI startup PrismML announced Bonsai 2 27B on September 17, the latest version of its ternary-quantized model family
Tag: PrismML
Break through the limits of hardware! PrismML compresses the 27B model to run locally on the iPhone, reducing the size by 90% but retaining nearly 90% of the strength
Although there are many small LLM models designed specifically for mobile phones, everyone who has used them knows that because ...
The volume is reduced by 14 times! PrismML launches the world’s first 1-bit LLM “Bonsai 8B”, which can run locally on iPhone
In the traditional impression, if you want an AI model to be smarter, you must make it larger, which means ...