# Cloudflare Trains Its First Models and Releases Them as Open-Weight Decision Engines

Clef is a 27-billion-parameter model built on a Qwen backbone that returns typed probabilities from a single prefill pass, with Cloudflare reporting a 38.8-millisecond median latency for the smaller Clef-flash.

- Published: 2026-10-03T06:33:24.401Z
- Canonical: https://polylog.news/ai/2026-10-03/cloudflare-trains-its-first-models-and-releases-them-as-open
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Polylog editors](https://polylog.news), [Cloudflare Blog](https://blog.cloudflare.com/clef-decision-models/), [MarkTechPost](https://www.marktechpost.com/2026/10/01/cloudflare-releases-clef-and-clef-flash/)

Cloudflare has released Clef and Clef-flash, its first in-house trained models, as open-weight decision models on Workers AI, alongside a reinforcement-learning fine-tuning service. The models do not generate text. Given an input and a set…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-10-03/cloudflare-trains-its-first-models-and-releases-them-as-open (subscription information: https://polylog.news/pricing).