Cloudflare's Clef decision models beat Jev on accuracy and latency, now open source

Clef: our open-source decision models

Cloudflare's Clef decision models beat Jev on accuracy and latency, now open source

Cloudflare released Clef and Clef-flash, two decision models that produce typed probabilistic outputs for agentic workflows. Clef leads the Jev Decision Index and is 2x faster than gpt-oss-120b in a threat intelligence test. The models are open-sourced under Apache 2.0 on Hugging Face and hosted on Workers AI. Cloudflare also launched an RL fine-tuning service to adapt Clef to specific use cases.

In music theory, a clef is a symbol placed at the beginning of a musical staff that assigns specific pitch names to the lines and spaces. A decision model is analogous to a music clef because it helps define the domain of the context and the subsequent notes (actions) that follow it.
  1. manlymuppet

    Am I hearing this right, that they made a decision model based on Typesafe's new paradigm, and actually made a model better than Jev based on Typesafe's own ranking?

    And it's only been a few weeks.

  2. vulture916

    Jev = $0.042/m input, output free

    Clef = $0.24/m input, no output price listed

    At 300 tokens per call, you'd get:

    One million decisions on Jev cost about $12.60.

    One million decisions on Clef cost about $72.

    Would probably make sense to self-host Clef, if you have the capability/resources. If not...

  3. buildbuildbuild

    Open weights, not open source.

    The weights have permissive licensing, but the data and training pipeline are not published to reproduce them from their proprietary Qwen starting points. Weights are not "source."

  4. ssiddharth

    Pricing is $0.24/million input tokens which is ~6x compared to Jev. Clef-flash is at $0.09 which is way more competitive.

  5. meander_water

    > This contrasts with the world of Large Language Models (LLMs), which are largely non-deterministic, but are open-ended enough to reason...

    This seems misleading. Decision models do not produce deterministic output. Repeated calls can product different decisions just like an LLM with structured outputs.

  6. bityard

    Clef is based on Qwen3.8-27B and Clef-flash is based on Qwen3.8-9B (edit: actually Qwen3.5-9B). So, similar in spirit to Kev by my understanding, but based on a newer model.

  7. mrkn1

    For smaller scale decision model that runs on CPU, check https://news.ycombinator.com/item?id=49923223

  8. amluto

    I’ll go out on a limb and suggest that I don’t think a Jev-like model is particularly useful unless you can fine tune it. The Jev API has zero ability to pass in a prior [0], and, if you can neither pass in a prior nor fine tune for your system, you will get an output that may be almost meaningless.

    I’d love to see someone build a model of this sort that can actually accept priors and do something intelligent with them.

    [0] You can feed Jev a prior as text. I’ve tried it. It works poorly.

More from this day

2026-10-01