Opus 4.6 Lands with Agent Teams and a $20 Billion Round

I'm Shannon, and this is the Claude Notes Brief -- your weekly rundown of Claude Code updates and Anthropic news for the week of February seventh. Opus four point six lands with experimental agent teams and automatic memory. Goldman Sachs deploys Claude-powered banking agents. And Apple brings agentic coding to Xcode.

Let's start with Claude Code -- seven releases this week, and the throughline is collaboration at scale. The headline is Claude Opus four point six, now available with two major new capabilities that fundamentally change how you can work with the tool. First, experimental agent teams -- multiple Claude instances that can collaborate in parallel on complex tasks. You enable this by setting the experimental agent teams environment variable, though be aware it's token-intensive since you're running multiple models simultaneously.

The second big addition is automatic memory -- Claude now records and recalls context as it works, reducing the need to re-explain your project setup across sessions. That collaborative theme extends to the tooling around agent teams too. If you're building multi-agent workflows, you now have three new capabilities: teammate idle and task completed hook events let you react to agent lifecycle changes, task syntax in frontmatter lets you control which sub-agents a given agent can spawn, and a new memory frontmatter field gives agents persistent memory scoped to user, project, or local levels. The model upgrade also brings fast mode support to Opus four point six, giving you lower-latency responses from the most capable model.

A related fix ensures the fast command becomes available immediately after enabling extra usage, rather than requiring a restart. Beyond the agent features, your session picker now shows git branch and message count for each session, and you can search by branch name to find the right context quickly. OAuth users can also browse and resume sessions directly from claude dot ai. PDF handling got upgraded too -- you can now read specific page ranges using the pages parameter in the Read tool, which helps avoid hitting size limits on large documents.

PDFs over ten pages now return a lightweight reference when mentioned rather than being dumped into context wholesale. And for those building with MCP servers, you can now pass client ID and client secret to the MCP add command for services like Slack that don't support Dynamic Client Registration. A new debug command also lets Claude help you diagnose session issues directly.

Under the hood, the focus was on security and memory efficiency. A security fix closes a gap where commands excluded from sandboxing could bypass the Bash ask permission rule when auto-allow was enabled -- important if you're using the dangerous disable sandbox setting. Several other fixes landed too: PDF size errors no longer permanently lock sessions, temperature override is now respected in the streaming path, and the prompt cache correctly invalidates when tool schemas change. The biggest performance win is a sixty-eight percent memory reduction for users with many sessions.

If you use resume frequently or have accumulated lots of past sessions, the tool now uses lightweight stat-based loading with progressive enrichment instead of loading the full session index upfront.

Turning to broader Anthropic news, it was a week of major partnerships and financial moves. Goldman Sachs is deploying Claude-powered agents to automate trade accounting and client onboarding, according to CNBC -- a concrete example of the agentic workflows that agent teams in Claude Code are designed to enable. Apple is bringing similar capabilities to Xcode, adding agents from both Anthropic and OpenAI to its coding tool. That expands where Claude Code's capabilities show up in developer workflows beyond the standalone tool.

On the financial side, Bloomberg is reporting that Anthropic's more than twenty billion dollar funding round is expected to close as soon as next week, alongside an employee tender offer at a three hundred fifty billion dollar valuation. The Wall Street Journal traces how Anthropic's focus on coding and enterprise clients drove what they're calling a market-moving week, providing useful background on the commercial trajectory behind the tools you use daily. That's it for the brief. I'm Shannon, and we'll see you next week.