Over the last few weeks, there has been lots of buzz around decision models such as Typesafe AI’s Jev System One model. Today, we’re releasing two Cloudflare-trained decision models, Clef and Clef-flash, hosted on Workers AI. Clef is currently the leader when evaluated against the Jev Decision Index. These models are smarter, faster, and fully Jev-API compatible. We’re fully open-sourcing these models on Hugging Face under an Apache 2.0 license. Lastly, we’re excited to debut our new reinforcement learning (RL) product, which allows customers to fine-tune Clef to suit their use cases.
A decision model makes classifications to help agents decide how to act, based on certain probabilities. For example, you can pass in a customer support message and ask if it is urgent and which team should handle it. A decision model will return typed answers with probabilities, which your code can use to route the ticket, trigger an escalation, or defer to a human.
At Cloudflare, we’ve been testing Clef on our Threat Intelligence team to classify website domains. This classification took 2.2s to fetch, render, and classify the website. In contrast, our fastest general LLM gpt-oss-120b took 4.7s in the same workflow. Clef has a vision encoder, a 64k context window, and scores competitively against other decision models.
Source: Hacker News · Summarized by HeadlinesBriefing