Tasks can now spawn async subagents that move to the background and continue working independently, even if the main agent finishes its task and becomes inactive.
This lets subagents handle long-running tasks, great for monitoring logs or waiting for builds.
Claude now compacts context exponentially faster.
Compacting takes only seconds so you don’t get interrupted.
You can also now rename sessions to make them easier to find & resume later.
Type /rename to give any previous session a custom name.
We’ve also added keyboard shortcuts to the /resume screen. Hit ‘R’ to rename a session or ‘P’ to preview a session.
Lastly, we’ve shipped a new command: /stats
/stats generates a visualization of your daily Claude Code usage and provides data on your sessions, usage streaks, and favorite models.
All features are available in today’s Claude Code build. Run `claude update` for the latest.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Claude Code's native installer is now generally available.
It's simpler, more stable, and doesn't require Node.js. We recommend this as the default installation method for all Claude Code users going forward.
What’s improved with the native installer:
- The auto-updater is now more stable: users should run into fewer failed updates and bricked installs
- Claude Code is a single, self-contained executable, and no longer has a Node.js dependency
If you’re already using Claude Code, you can migrate to the new installer:
- Close all Claude Code sessions
- Run `claude install` in your terminal
For Alpine and musl/uClibc-based distributions, see our full docs below before updating.
Introducing Claude Haiku 4.5: our latest small model.
Five months ago, Claude Sonnet 4 was state-of-the-art. Today, Haiku 4.5 matches its coding performance at one-third the cost and more than twice the speed.
Haiku 4.5 surpasses Sonnet 4 on computer use tasks, making Claude for Chrome even faster.
In Claude Code, it makes multi-agent projects and rapid prototyping markedly more responsive.
Sonnet 4.5 remains the best coding model in the world. Haiku 4.5 gives you near-frontier performance with greater cost-efficiency.
Use them together: Sonnet can build multi-step plans, then orchestrate a team of Haikus to complete subtasks in parallel.