Cloudflare

Cloudflare's decision model now reads audio and video, and the cheap tier fell to $0.038 with a smaller window

The new model is built on Alibaba's Qwen3-Omni with the speech output stripped out. Cloudflare says 0.24% of requests exceed the shorter window.

Microsoft's first decision model is a post-trained Qwen, costs $0.042 a million input tokens and charges nothing for output

The model is built on Alibaba's open Qwen3.5-9B, and Microsoft says it will later rebase the same system on its own MAI models and on OpenAI's.

The Deno team is going to Cloudflare, the runtime gets one more year of fixes, and Ryan Dahl says the draw is agent infrastructure

Dahl is folding his team into the Workers and Durable Objects groups, and names persistent state and cheap serverless execution as what he wants for agent harnesses.

Cloudflare has opened two models that never write a word, and answers a typed question in 39 milliseconds

The smaller model answers in a median 38.8ms against Jev's 524.1ms on Cloudflare's own 43-benchmark run — and all of the numbers so far are the vendor's.

Cloudflare has turned on a paywall that bills the agent instead of asking it to log in

The closed beta is open to US sellers and buyers only. Four products are already charging through it, including Cloudflare's own AI Gateway.

Cloudflare will bill AI companies when they use a publisher's work, not when they crawl it

The design puts the meter on the AI company's side: usage is self-reported through a single API, and Cloudflare says it verifies the reports. No buyers were named at launch.