PrismML launches 4GB Bonsai 27B AI model that can run on a phone
PrismML has launched Bonsai 27B, a compressed AI model based on Qwen3.6 27B. The 1-bit version uses about 3.9GB of memory and is built to run locally on smartphones and laptops. The model supports coding, reasoning, image understanding and tool use, though benchmark scores show a clear drop compared with the full-precision version.
Why this byte is shareable
Signal quality
observed
Confidence badge and source context included.
Entity anchor
Open Source
Clear company or model context for distribution.
Export ready
1200 x 630 card
Optimized for X, LinkedIn, and chat previews.
Why it matters
Agent runtime/tool-calling changes can shift schema behavior and fallback paths. Validate tool-call success and retries before shipping.
Suggested launch post
Use this in X threads, community posts, internal team chats, or launch recaps.
PrismML launches 4GB Bonsai 27B AI model that can run on a phone Why it matters: Agent runtime/tool-calling changes can shift schema behavior and fallback paths. Validate tool-call success and retries before shipping. Source: News9live https://a2zai.ai/bytes/prismml-launches...
Permalink: https://a2zai.ai/bytes/prismml-launches-4gb-bonsai-27b-ai-model-that-can-run-on-a-phone-e1796491
Social card: https://a2zai.ai/bytes/prismml-launches-4gb-bonsai-27b-ai-model-that-can-run-on-a-phone-e1796491/opengraph-image