Clef-flash
A lightweight 9-billion-parameter decision model released by Cloudflare, with a median decision latency of 38.8 milliseconds. It is deployed on Workers AI edge nodes and targets high-frequency, low-latency real-time decision scenarios.
Timeline (last 90 days)
Cloudflare Clef / Clef-flash and AWS Strands Decider 2B were both released on October 1, with undisclosed prices; Cloudflare is an open-source model.
clef-omni is the third model in a family only ten days old; Cloudflare released Clef and Clef-flash on October 1 as an open-weights response to Jev; Jev is TypeSafe's decision model, released on September 15.
Eight days later, Cloudflare cut Clef-flash from $0.09 to $0.038 (already below Jev), migrated Clef's inference to SGLang for a 1.7 to 2.0 times median speedup, and released clef-omni; Clef itself still remains at $0.24.
The price of Clef-flash is approximately three-quarters lower than that of the standard version.
The lightweight Clef-flash is a 9B parameter model, priced at $0.09 per million input tokens.
Both the Clef and Clef-flash models claim compatibility with Jev.
Clef-flash has a median decision latency of only 38.8 milliseconds.
All Facts (7)
Cloudflare Clef / Clef-flash and AWS Strands Decider 2B were both released on October 1, with undisclosed prices; Cloudflare is an open-source model.
50%UnverifiedThe price of Clef-flash is approximately three-quarters lower than that of the standard version.
50%UnverifiedBoth the Clef and Clef-flash models claim compatibility with Jev.
50%UnverifiedThe lightweight Clef-flash is a 9B parameter model, priced at $0.09 per million input tokens.
50%UnverifiedClef-flash has a median decision latency of only 38.8 milliseconds.
50%Unverifiedclef-omni is the third model in a family only ten days old; Cloudflare released Clef and Clef-flash on October 1 as an open-weights response to Jev; Jev is TypeSafe's decision model, released on September 15.
50%UnverifiedEight days later, Cloudflare cut Clef-flash from $0.09 to $0.038 (already below Jev), migrated Clef's inference to SGLang for a 1.7 to 2.0 times median speedup, and released clef-omni; Clef itself still remains at $0.24.
50%