Claude Code v2.1.267 Update: Prompt Cache Optimization and Stability Improvements

Claude Code released v2.1.267 with per-model max effort level settings and multiple prompt cache reuse bug fixes. Issues resolved include cache invalidation on

Claude Code has released version v2.1.267. Key improvements include the addition of per-model maximum effort level settings and fixes for multiple bugs related to prompt cache reuse. Issues resolved include cache invalidation caused by tool list rewriting on MCP server reconnection, tool definitions being resent on session resume, missing parallel tool call data when resuming large sessions (5MB+), and improved retry logic for AWS/Google Cloud authentication expiration.

Claude Code · Release v2.1.267 · 2026-09-09

weltiq perspective: Prompt cache optimization reduces API call costs and latency during backtesting and automated execution, improving strategy testing efficiency for weltiq users.

Release Notes Excerpt

What's changed Added maxEffortLevel setting (top-level or per model under modelSettings ): caps the effort level on every provider, including Bedrock, Vertex and Foundry; users can still pick a lower level Added --system-prompt-snapshot off to render the system prompt fresh on every request instead of reusing the conversation's recorded prompt (for iterating on prompt text) Fixed Cowork scheduled tasks in the cloud failing at startup for organizations whose managed settings require sandboxing Fixed /context and other local command output rendering blank on mobile clients Fixed shift+enter and option+backspace not working after reconnecting to a tmux or ssh session inside an agent view Fixed …

Source: https://github.com/anthropics/claude-code/releases/tag/v2.1.267