Dispatch
Cloudflare Workers AI Runs Frontier Open-Source Models Starting with Kimi K2.5 — 77% Cost Reduction
Cloudflare announced large model support on Workers AI with Kimi K2.5 (256k context, vision, multi-turn tool calling) as the first offering, engineered specifically for agent workloads. Internal testing showed switching a 7-billion-token-per-day security code review agent from a proprietary mid-tier model to Kimi K2.5 cut annual costs from ~$2.4M to a fraction. New prefix caching with session affinity headers and redesigned async APIs address the core agent scaling bottleneck: cold token cost at volume.
↳ Follow the thread