18 August 2026

Anthropic releases cost-cutting feature for Claude, discloses security breach

• Anthropic published guidance on prompt caching, a technique that reduces repeated input costs to 10 percent for Claude Code users.
• The company is testing a side-by-side interface letting users compare Claude's performance against other models directly.
• Anthropic disclosed that safety filters for biological and chemical weapon requests were accidentally disabled for nearly a year, exposing approximately 133 million contractor interactions.

How it was covered

AI BreakfastIndependent editors

Anthropic released a guide for reducing Claude Code token costs through prompt caching, which cuts repeat input costs to 10 percent, and is testing a side-by-side model comparison interface. The newsletter emphasizes that an internal model outperforms Mythos 5, but highlights a major safety issue where 133 million contractor requests were exposed after biological and chemical weapon classifiers were accidentally left off for nearly a year.