Cloudflare has released Clef and Clef-flash, two decision models hosted on its Workers AI service and available under an Apache 2.0 license. They return structured choices and probabilities, such as whether a support request is urgent and which team should handle it. Cloudflare says the models can also classify images and accept up to 64,000 tokens of context.
The company says Clef is available through Workers AI and on Hugging Face. It is also starting a hands-on fine-tuning service with its forward-deployed engineering team; a self-serve platform is planned for later, with no date given. Cloudflare reports Clef took 2.2 seconds to fetch, render and classify a website in an internal test, compared with 4.7 seconds for its fastest general model in that workflow.
Why it matters
Developers building workflows that route support requests or classify websites can try Clef through Workers AI or run the open-source models themselves; how well they perform on a particular workload remains to be tested.
