Cloudflare’s Clef models return decisions instead of chatbot text

Decision Models

Cloudflare released two open-source models built to return labels or choices instead of open-ended text. That targets routing and classification jobs where a general-purpose LLM can be overkill: image classification, bug routing, and sending support tickets to the right team.

Clef has 27 billion parameters and Clef-flash has 9 billion. Both run on Workers AI, with weights on Hugging Face and local access through Ollama after its October 3 announcement. Cloudflare says the models are designed for high-speed classification and agentic workflows, making the pitch less “another chatbot” and more a fast decision layer.