How to cut response latency in your Claude API app with /claude-api latency-optimize in Claude Code

Claude Code 2.1.319 (November 6) added /claude-api latency-optimize. It finds where your Claude API code makes people wait and tries faster settings against your eval. Use it on a chatbot or agent app whose users say responses are slow.

1. Check that Claude Code is 2.1.319 or later.

claude --version

1. If you don't have an eval set yet, build one first. build-eval has been included since 2.1.259.

/claude-api build-eval

1. Open Claude Code in your project folder and type the command.

/claude-api latency-optimize

For cost, use /claude-api cost-optimize (2.1.247 or later). It profiles where your Claude API spend goes and proposes savings from options such as prompt caching, trimming unneeded input and output tokens, batch processing, effort, and model choice, one change at a time. latency-optimize isn't in the skills docs' subcommand table yet; it appears only in the changelog.