█

AI/TLDR

Anthropic · 2026-10-26 · notable

Claude Code 2.1.311 — ultrathink now runs a turn at xhigh effort

Claude Code 2.1.311 makes ultrathink run its turn at xhigh effort on Opus 5.5, Sonnet 5.5, Haiku 5.5 and Fable 5.1. It also stops standard-context sessions from halting past 200K tokens and fixes dozens of bugs.

Claude Code repository card on GitHub

Claude Code 2.1.311 ties ultrathink to xhigh effort and fixes a long list of session, MCP and subagent bugs.

What is it?

Typing ultrathink in the Claude Code 2.1.311 terminal now runs that turn at xhigh effort on Opus 5.5, Sonnet 5.5, Haiku 5.5 and Fable 5.1 through Anthropic's API. Typed while Claude is already working, the keyword raises the effort for the rest of that turn.

How does it work?

The self-hosted runner gains --hold-idle-file and --max-hold-min flags, so an idle session is kept while the host reports processes it left running. Sessions with a standard context that go over 200K tokens no longer stop with a 'Usage credits required for 1M context' error; the conversation is compacted and work continues.

Why does it matter?

Most of the release is reliability work for long, agent-heavy sessions. Subagents can now use MCP tools from servers that connect after they start, SubagentStop hooks fire once instead of about twice a minute, and Write and Edit no longer stall for seconds on very large files.

Who is it for?

developers using Claude Code with subagents, MCP servers or self-hosted runners

Try it

npm i -g @anthropic-ai/claude-code@2.1.311

Sources

Tags

  • claude-code
  • anthropic
  • coding-agent
  • cli
  • ultrathink
  • effort
  • self-hosted-runner
  • compaction
  • subagents
  • mcp
  • developer-tools
  • release

← All releases · Learn AI