PrismML has released a compressed reasoning language model designed to bring advanced AI capabilities to PCs and smartphones. Its approach could reduce reliance on large cloud-based systems by making powerful models smaller, cheaper and more accessible locally.

The company says Bonsai 2 27B compresses Alibaba’s Qwen3.8 27B model to 5.9GB, reducing memory requirements by roughly nine to ten times while retaining 98% of benchmark performance. PrismML’s earlier Bonsai model reportedly achieved 95% parity and more than 11 million downloads, while its smaller models received a further 2.6 million; the startup has raised $22.25 million and plans to compress models with several hundred billion parameters.