Anthropic’s Claude team announced on X that “fast mode” for Claude Opus 4.7 is now in research preview on the API and in Claude Code. For Claude Code users, the mode is “opt-in today” and will become the default on Thursday, according to the company’s posts.
The company also listed several other products where Opus 4.7 fast mode is available in research preview, including Cursor, Emergent, Factory, v0, Warp, and Windsurf. For API users interested in trying it, ClaudeDevs pointed to a waitlist.
The announcement drew a familiar set of questions from developers. Some commenters asked whether fast mode uses a different inference architecture or simply optimized serving, and whether it changes quality on more complex reasoning tasks. Others focused on billing and limits, asking whether the mode increases usage consumption, whether it will land in the desktop app, and whether existing plans include it.
A number of replies treated the update as a latency improvement that could make longer workflows more practical, while others worried it would make users hit rate limits sooner. One commenter asked whether faster inference comes with a “usage multiplier,” while another noted that “the speed feels crazy good” after testing it.
Source: ClaudeDevs
