Model overview
What this model is for
bonsai-2:27b is a 27-billion-parameter vision model from Prism ML, available through Ollama, and is intended for applications that process image inputs. It belongs to the Bonsai 2 family, which is advertised as using ternary weights to produce smaller models with quality close to full-precision models. The supplied model record lists image modality support and tool-calling capability, so practical uses include vision workflows and applications where the model needs to invoke tools as part of a task. The supplied material does not provide an advertised context-window size, inference-speed benchmark, pricing, or licensing terms, so those factors should be verified before deployment. It also does not describe specific supported image formats, resolution limits, hardware requirements, fine-tuning options, or evaluation results. The model page identifies the exact variant as bonsai-2:27b and links to its Ollama library entry. Developers should therefore treat the quality, memory footprint, tool-calling behavior, and operational requirements as deployment questions requiring local validation rather than assuming them from the family description alone.