Cloudflare Open Sources Decision Models for AI Agents

During its recent "Birthday Week", Cloudflare announced Clef, a set of open-weight AI models designed to choose between predefined options rather than generate text. Cloudflare released 9B- and 27B-parameter models, along with a platform for adapting them to specific decision-making tasks.

Clef is a 27B multimodal model that takes a state and a schema of typed questions as input and returns decisions. It can process text, JSON, images, or video, and returns a probability for each allowed option for every question in a single forward pass, without generating free-form text or requiring output parsing.

Source: Cloudflare blog

A decision model classifies inputs and returns typed outcomes with probabilities, allowing AI agents to use these results to determine how to act, such as routing a support request, escalating it, or deferring the decision to a human. Cloudflare’s API is compatible with the popular Typesafe AI’s Jev System One model.

Source: Cloudflare blog

Michelle Chen, group product manager at Cloudflare, Alex Reneau, principal machine learning engineer at Cloudflare, and Kevin Flansburg, senior engineering manager at Cloudflare, write:

Because they are hosted on Cloudflare’s infrastructure, we’re able to take advantage of our GPUs at the edge, leading to low network latency and faster decisions. This means that you could put Clef into the hot path for agents to make decisions and combine that with one of our LLMs on Workers AI to take action.