Clef-Flash

Name in the API

Cloudflare/clef-flash

Base model

Cloudflare/clef-flash

Type

Decisions (classification)

Input → output

text or JSON, plus typed questions → a probability for every allowed answer

Context length

8,192 tokens

Availability

Dedicated

Hardware

A 24 GB slice of a GPU per instance

Limits

Text only. Images, audio and video are not accepted.

At most 64 questions per request, 26 options per choice question and 10 levels per score question.

An input longer than 8,192 tokens is rejected with error 400. If only the questions push it over the limit, the input is shortened.

Endpoint

/v1/systemone, /v1/decisions

Example

Decision models

Notes

• Clef-Flash does not write text. It answers several questions about one input in a single pass and returns the probability of every allowed answer: one of several options (choice), yes or no (predicate in /v1/decisions, noul in /v1/systemone) or a level on a scale (score).