ainewsblitz.com

Breaking

Claude Code Users Can Extend Subagent Prompt Cache to One Hour, Reducing Token Costs

  • Foundation Models
  • AI Agents
  • Sales & Business

Claude Code users can now extend their subagent prompt cache lifetime from the default five minutes to one hour by configuring a single setting, a move that can materially reduce token consumption for long-running workflows. By default, Claude Code's main conversation gets a one-hour prompt cache on Claude subscriptions, but subagents, workflows, and background requests expire after five minutes—causing the context to recompute and re-read on the next call. A simple configuration change in ~/.claude/settings.json allows users to match subagents to the main conversation's caching behavior.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 26,559 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year