Claude Opus 4.6 is Claude’s most capable model yet, built for deep reasoning, long-running agentic tasks, and large codebases. With a 1M token context window, adaptive thinking, and improved planning, it delivers state-of-the-art performance across coding, analysis, research, and real-world work.
Hey everyone 👋
Sharing Claude Opus 4.6, Claude’s newest and most capable model. Built for large codebases, long-running agent workflows, and deep reasoning. It handles huge context, plans before acting, and stays reliable over time.
Would love your thoughts.
I feel like this is the first time my AI assistant has a really sharp and witty sense of humor and I love it. Opus 4.6 might be my new favorite.
I really love that Claude is focusing on thinking more instead of a speed. I am using Claude Code constantly now, and FEEL how it gets smarter, 4.6 is top of the top. Thank you so much guys for making my project be faster than ever ❤️
The launch frequency is getting out of hand - building faster than we can even look at the products haha! Will poke around later today but congrats on the rollout
Agent Teams sound incredible! How do you manage token costs when the lead agent synthesizes results from multiple sub-agents? Seems like coordination overhead could add up fast.
Excited to try this!
1M context + actually plans before acting is the hook for me. My monorepo’s a tangle and agents drift by hour 2. If this keeps focus and finishes, not just chatters, that’s huge. I’ll throw it a week‑long cleanup and see if it stays on the rails.
It has topped the tests again! What is its most distinctive feature that sets it apart from the others?
So excited to be integrating it on Krater, a lot of our users were looking forward to it!
Hi Team, congrats! What are the best use cases for financial teams you can see now?
Huge congrats on the launch — Opus 4.6 looks like a beast for deep, long-horizon coding work.
Perhaps the most interesting aspect of 4.6 is buried here:
We’ve introduced agent teams in Claude Code as a research preview. You can now spin up multiple agents that work in parallel as a team and coordinate autonomously—best for tasks that split into independent, read-heavy work like codebase reviews. You can take over any subagent directly using Shift+Up/Down or tmux.
Super interesting:
Agent teams let you coordinate multiple Claude Code instances working together. One session acts as the team lead, coordinating work, assigning tasks, and synthesizing results. Teammates work independently, each in its own context window, and communicate directly with each other.
Unlike subagents, which run within a single session and can only report back to the main agent, you can also interact with individual teammates directly without going through the lead.
