Opus 5.5 fast mode goes GA at $8/$40 per million tokens

Anthropic's developer account says fast mode for Opus 5.5 is no longer a research preview and the waitlist is gone. The model is unchanged; the difference is speed, up to 2.5x faster, at $8 per million input tokens and $40 per million output tokens. Fast mode is also available through Microsoft Foundry hosted on Azure, with Google Cloud listed as coming soon. Claude Code's own fast mode is not affected and still draws on extra usage credits. The change matters most for agent loops where latency compounds across many sequential calls. A faster turn shortens wall-clock time for long tasks, but the per-token price is the cost of that speed, so teams should measure whether the time saved is worth it on their own workloads rather than switching every call. The announcement gives no benchmark of task completion time, only the speed multiple.