Clef
Cloudflare's Apache-2.0 27B multimodal decision model that scores every option of every question jointly, with a Jev-compatible API.
Overview
Clef is a 27B multimodal decision model from Cloudflare, post-trained from Qwen3.8-27B. It turns a state and a schema of typed questions into decisions, reading the state as text, JSON, images or video and returning a probability for every allowed option of every question in a single forward pass, with no free-form generation or output parsing.
Its architecture adds a small joint schema head on the backbone's final hidden states that routes evidence from the state to each question and scores all options jointly. The API is fully compatible with Jev and SystemOne, served at /v1/systemone. Clef-Flash is a smaller, faster variant. Both are hosted on Cloudflare Workers AI, and the launch post reports Decision Index results with a public leaderboard. Weights are Apache-2.0 on Hugging Face.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does this message express urgency? / Is the total above 1000 USD? | yes · no |
| Choice | Any user-defined choice over the state | — |
| Score | Any user-defined ordered score | — |
How it works
- Send state (text, JSON, image or video) and typed questions
- Backbone encodes the state
- Joint schema head scores all options of all questions
- Return Jev-shaped probabilities
Strengths
- Multimodal state input
- Open weights, Apache-2.0
- Drop-in Jev API compatibility; hosted on Workers AI
Caveats
- 27B model needs substantial hardware to self-host
- Custom code in the model repo
Getting started
Call /v1/systemone on a Clef server or Workers AI with model Cloudflare/clef.At a glance
- Type
- Service
- Status
- production
- Language
- Python
- Runtime
- transformers / Workers AI
- Has code
- yes
- License
- Apache-2.0
Use cases and integrations
How curated lists describe it
Cloudflare's Apache-2.0 decision models post-trained from Qwen3.8-27B: Clef reads text, JSON, images, or video and scores every option of every question jointly, and Clef-flash is the smaller, faster variant. Both are hosted on Workers AI with a Jev-compatible API. Launch post with Jev Decision Index results: [blog](https://blog.cloudflare.com/clef-decision-models).
AnotiaWang/awesome-decision-models
Frequently asked questions
Is it compatible with Jev clients?
Yes, the API is fully compatible with Jev and SystemOne.
Can it read images?
Yes, text, JSON, images or video.
Is there a smaller version?
Yes, Clef-Flash.
Related projects
Free hosted Jev-style multimodal decision model API (about 20B active) that picks actions from images and options, aimed at robot control.
ServicebetaComfyUI nodes for Jev-style noul, choice and score decisions over images and text using a local imajev adapter on Qwen3.5.
ProjectbetaPythonResearch code for an image-and-text System One model that reads option-letter logits from a multimodal LLM, with training and calibration.
ProjectresearchPythonIndependent open model on Molmo2-ER that answers spatial questions by choosing from fixed options with one shared classifier head.
ProjectresearchPythonOpen project training a vision-enabled Jev-like model (0.8B, 9B) that judges dynamic candidate sets over images with Choice, Noul and Score.
ProjectresearchPythonIndependent Jev-style server on DiffusionGemma and SGLang that classifies doodles, flowers and messages with Choice, Score and Noul.
ServiceresearchPython