Tag
4 articles
PrismML has brought its compact language models to Qualcomm-powered smart glasses, enabling on-device AI processing that enhances privacy and reduces latency.
PrismML has released Ternary Bonsai 2 27B, a 5.9 GB ternary-weight model that retains 98.2% of Qwen3.8 27B’s performance. The model supports text and image inputs with a 262K-token context window.
Developers have successfully deployed the 1-bit Bonsai-27B language model using PrismML’s fork of llama.cpp, enabling efficient local inference with OpenAI-compatible workflows.
PrismML has compressed a 27-billion-parameter AI model to under 4 GB, enabling it to run on an iPhone. Apple is reportedly testing the technology, which could advance on-device AI capabilities.