All models
Cloudflare
Clef-flash
A smaller Cloudflare decision model for frequent routing and classification tasks, with a lower input-token price than Clef.
Plain-English overview
What Clef-flash actually is
Clef-flash uses a 9B backbone and the same typed question format as Clef. It is aimed at workflows where a quick, bounded answer is more useful than a generated paragraph.
Use it to try lower-cost decisions in a queue or before an agent action, then compare errors against Clef and Jev on your own data. Workers AI supports embedded images; video-frame inputs are documented for self-hosting.
Good fit for
- Support ticket classification and routing
- High-volume structured decisions
- Agent decisions with human escalation
Category comparison
The facts that matter for decision models
These are provider-published specifications, not Cody benchmark scores. Follow the linked sources for current limits and endpoint-specific exceptions.
- Input price
- $0.038 / 1M input tokensCurrent provider price for the text or structured state submitted to the model, including whether decision outputs are billed.
- Input capacity
- 24,576 tokens on Workers AI; up to 64 questions per requestMaximum documented state and question capacity for one request.
- Decision types
- Choice, Score, and Noul probability questionsThe bounded question forms the model can answer, such as choices, scores, or probabilities.
- Output contract
- Typed JSON decisions defined before inferenceHow the response is constrained for direct use by application code.
- Uncertainty
- Probability distributions; Choice and Score also include confidenceWhether the endpoint exposes probabilities, confidence, or another calibrated signal.
- Where to use it
- Workers AI API and self-hosted weightsCurrent API, SDK, playground, and access-stage information.
API and provider access
Where to get Clef-flash
Availability
Regions and access stage
Available through Workers AI and as downloadable Apache 2.0 weights for self-hosting.
Workers AI runs on Cloudflare's network. The reviewed Clef documentation does not specify country-by-country access or a selectable data-residency region; confirm requirements with Cloudflare.
Check live availabilityData and training
The route matters.
Cloudflare says Workers AI customer content is not used for training or service improvement without explicit consent. Connected storage services can retain content. Self-hosted deployments follow your own storage and access controls.
This is a concise reading of the cited provider material, not legal advice. A third-party gateway can have different storage, routing, training, and residency terms from the model maker's direct API.
Read the provider policyFrequently asked questions
Clef-flash FAQ
What is Clef-flash?
A smaller Cloudflare decision model for frequent routing and classification tasks, with a lower input-token price than Clef. Clef-flash uses a 9B backbone and the same typed question format as Clef. It is aimed at workflows where a quick, bounded answer is more useful than a generated paragraph.
When was Clef-flash released?
Clef-flash was released on October 1, 2026 according to the cited provider materials.
Where can I access Clef-flash?
Available through Workers AI and as downloadable Apache 2.0 weights for self-hosting. The access routes listed in this guide are Cloudflare.
How much does Clef-flash cost?
$0.038 / 1M input tokens. Workers AI lists input-token pricing only, billed through Neurons. Free daily allowances and other Cloudflare services are separate; self-hosting has its own infrastructure costs.
Where is Clef-flash available?
Available through Workers AI and as downloadable Apache 2.0 weights for self-hosting. Workers AI runs on Cloudflare's network. The reviewed Clef documentation does not specify country-by-country access or a selectable data-residency region; confirm requirements with Cloudflare.
Is my Clef-flash API data used for training?
Cloudflare says Workers AI customer content is not used for training or service improvement without explicit consent. Connected storage services can retain content. Self-hosted deployments follow your own storage and access controls. The policy belongs to the provider route and account terms, so verify it again before production use.
Related comparisons
Head-to-head comparisons
Clef-flash vs Clef
Cloudflare's larger decision model for classifying information and choosing an action without generating a chat response.
Open full comparisonClef-flash vs Jev 1.13
TypeSafe AI's first System One model for fast, typed choices, scores, and probabilities that software can use directly.
Open full comparison
