Using Clef-flash on Cloudflare Workers AI

Implementation guide · Clef · Cloudflare

ServerlessOpen Source

Cloudflare Workers AI exposes Clef-flash through model ID @cf/cloudflare/clef-flash. Use the setup steps, sourced pricing, capabilities, and official provider links below to validate this route before deployment.

Last refreshed 2026-10-04. Next refresh: weekly.

Quick Start

  1. 1
    Create an account at Cloudflare Workers AI and generate an API key.
  2. 2
    Use the Cloudflare Workers AI SDK or REST API to call @cf/cloudflare/clef-flash — see the documentation for request format.
  3. 3
    You'll be billed $0.09/1M input. See full pricing.
API PortalDocumentationPricingModel Card

Code Examples

See Cloudflare Workers AI documentation for integration details.

Pricing on Cloudflare Workers AI

TypePrice (per 1M)
Input tokens$0.09

Capabilities

VisionMultimodalStructured Outputs

About Clef-flash

Clef-flash is Cloudflare's 9B open-weight multimodal decision model, the fast sibling of Clef, released 2026-10-01 by the Workers AI team. Post-trained from Qwen3.5-9B (not Qwen3.8): the backbone is frozen, and a joint schema head is trained alongside rank-256 LoRA adapters with label-smoothed cross-entropy, a Brier calibration loss, and RLCD. It reads a state (text, JSON, or images) and up to 64 typed questions (noul, choice, or score) and returns a probability for every allowed option in one non-autoregressive pass, with no free-form text generation.

Full model details →

Model Specs

Released2026-10-01
Parameters9B
Context66k

More Models on Cloudflare Workers AI

Clef
All models on Cloudflare Workers AI →

Provider

Cloudflare Workers AI

Cloudflare

San Francisco, California, United States