Claude, the essentials — edition of November 16, 2026
Anthropic's IPO roadshow, self-hosted agents and a new default Fable model
Anthropic has set a $70 to $78 price range for its proposed Nasdaq listing, while the Claude platform adds a fully self-hosted Managed Agents option and Claude Code ships a new default model. Three status-page incidents were also reported in the same window.
- Anthropic launched the roadshow for its proposed IPO: 825,000,000 Class A shares at an expected $70 to $78, with a Nasdaq application under the symbol “ANT”. The registration statement has not yet become effective.
- Self-hosted Claude Managed Agents lets customers run the whole service, and keep agents, sessions and session history, on their own infrastructure.
- Claude Code 2.1.321 made Claude Fable 5.2 the default Fable model, with a 1M context window. Later releases added the experimental Code Mode, plugin version constraints and a zstd security fix.
- The status page logged an ongoing Sign in with Apple problem on iOS, a resolved purchasing incident, and a resolved broad error spike across claude.ai, the API and Claude Code.
Anthropic takes its public offering to investors
Anthropic, PBC announced on November 10 that it had launched the roadshow for its proposed initial public offering, following the S-1 it filed with the SEC on October 26. The offering covers 825,000,000 shares of Class A common stock. Anthropic is selling 801,351,354 of them and certain existing stockholders are selling 23,648,646. Anthropic says it will not receive any proceeds from the selling stockholders' shares. The underwriters also have a 30-day option to buy up to 123,750,000 additional shares.
The expected price range is $70.00 to $78.00 per share, and Anthropic has applied to list on the Nasdaq Global Select Market under “ANT”. The announcement is explicit about its limits. A registration statement has been filed but has not yet become effective, the shares cannot be sold or offers accepted before it does, and the notice is neither an offer nor a solicitation. It names Goldman Sachs & Co. LLC and Morgan Stanley & Co. LLC as sources for the preliminary prospectus, which is the only basis on which the offering will be made.
Sources: Anthropic launches roadshow for proposed initial public offering — Anthropic
Platform and Claude Code: more control, a new default model
The release notes for November 12 describe self-hosted Claude Managed Agents. Customers run the entire Managed Agents service on their own infrastructure, not just self-hosted sandboxes, and they store agents, sessions and session history there. The server sends model requests to the provider the customer chooses. In practice this moves the control plane and the data it accumulates inside the customer's perimeter.
Claude Code's recent releases point the same way. Version 2.1.321 added Claude Fable 5.2 (claude-fable-5-2) as the default Fable model, with a 1M context window, a price of $8/$40 per million tokens and cache reads at $0.20 per million. Version 2.1.322 added fast mode on Vertex in the global, us and eu regions, falling back to standard speed if Vertex refuses it. Version 2.1.323 introduced Code Mode, an experimental opt-in in which Claude writes JavaScript to call its tools. The documentation covers turning it on, or off for an entire organization. Plugin authors also get a version gate: with engines["claude-code"] in plugin.json, plugins outside their stated version range are not installed, updated or loaded, and install and update pick the newest version that fits. A TypeScript type check of a plugin's hooks module was added to claude plugin validate, where findings are warnings and --strict makes them fail.
Version 2.1.325 fixed a security issue in how zstd-compressed responses are decoded. The release notes give no further detail. It also added a hint under the error shown when an organization's policy blocks a request, telling users to run /rewind and change the message, or run /clear, and a spinner tip about re-running a notebook cell after Claude has read or edited a Jupyter notebook.
Sources: Self-hosted Claude Managed Agents (API release notes, November 12, 2026) — Anthropic (release notes) · Claude Code 2.1.321 — Claude Fable 5.2 — Claude Code · Claude Code 2.1.323 — Code Mode and plugin engines — Claude Code · Claude Code 2.1.325 — zstd security fix — Claude Code
Reliability: three incidents on different surfaces
The status page reports three separate problems, each on a different layer of the service. The first is still open. Sign in with Apple on the Claude iOS app is failing for some users, mainly new accounts. Anthropic is still investigating and says new users can sign up with Google or email instead.
The second is resolved. Purchases and plan changes on claude.ai had elevated errors from 18:34 PT on November 13 (02:34 UTC on November 14) to 20:03 PT (04:03 UTC), roughly an hour and a half. Anthropic says delayed plan changes and payments have now been applied. The third is a broader spike of errors on claude.ai, including the mobile apps, the Claude API and Claude Code. Errors are back to normal levels and Anthropic is continuing to monitor. Existing connectors such as Docs and Sheets work again, and Claude Managed Agents and Claude Meetings are recovering. Adding a new connector on claude.ai is temporarily unavailable, and the notice also mentions scheduled tasks and routines.
Sources: Some users unable to sign up with Apple on the Claude iOS app — status.claude.com · Elevated purchasing errors on claude.ai — status.claude.com · Elevated errors on claude.ai, the Claude API, and Claude Code — status.claude.com
New Anthropic publications on safety, multi-agent teams and hardware
Three further Anthropic publications are on the list, but the material available to us for each is little more than its title, so we do not describe their contents. One is “Building strong safety evaluations.” Another, from the Alignment Science blog, is “Worse Together: How Performance Breaks Down in Multi-User Multi-Agent Teams.” The third is “Claude can design electronic prototypes from start to finish.”
Read together, the titles suggest three directions: how to measure model safety, how agents behave when several people and several agents share a task, and how far Claude can carry a physical design. Readers who want the findings should go to the original pages.
Sources: Building strong safety evaluations — Anthropic · Worse Together: How Performance Breaks Down in Multi-User Multi-Agent Teams — Anthropic Alignment Science blog · Claude can design electronic prototypes from start to finish — Anthropic
This edition is an original synthesis written by Claude from aggregated news — Anthropic's own sources first (release notes, status, newsroom, research, engineering), then the press, Hacker News, Reddit and GitHub, under the editorial supervision of Héra SASU. Every fact links to its article, publisher named. See the live feed →
Claude News is published by Héra SASU. Independent media, not affiliated with Anthropic.