Claude News
FR
Live

Claude, the essentials — edition of October 1, 2026

Claude Sonnet 5.5 becomes the default Sonnet as Sonnet 4.5 gets a retirement date

By ·

Anthropic's Sonnet 5.5 arrives with a 1M-token context and a published price list. Sonnet 4.5 is scheduled to leave the Claude API on November 30, 2026. The day also brought a status-page incident that touched sign-in and several Claude products.

Key points

Sonnet 5.5 takes over the default slot

Anthropic introduced Claude Sonnet 5.5 and published its system card at the same time. A developer guide titled "Building with the Claude 5.5 Family: Choosing the Right Model and Getting More from Every Token" is also in the day's items. The concrete terms come from the Claude Code 2.1.284 release note. Sonnet 5.5 (claude-sonnet-5-5) is now the default Sonnet model on the Anthropic API. It offers a 1M-token context and is priced at $2/$10 per Mtok, with cache reads at $0.20/Mtok.

The items for the announcement, the system card and the guide carry only their titles, so this edition does not describe their contents. The titles frame the release around two things: choosing the right model within the 5.5 family, and efficient use of tokens. The release note's pricing, with a separate rate for cache reads, is stated plainly.

Sources: Introducing Claude Sonnet 5.5 — Anthropic · Claude Sonnet 5.5 System Card — Anthropic · Building with the Claude 5.5 Family: Choosing the Right Model and Getting More from Every Token — Anthropic · Claude Code 2.1.284 release notes — Anthropic (Claude Code)

Operational notes: a retirement date and an incident

The API release notes, in an entry dated September 30, 2026, announce the deprecation of Claude Sonnet 4.5 (claude-sonnet-4-5-20250929). Retirement on the Claude API is scheduled for November 30, 2026, and Anthropic recommends migrating to Claude Sonnet 5.5. The new model is therefore both the default Sonnet and the stated destination for users of the older one. The notice gives them about two months to move.

On the reliability side, the official status page records an incident of elevated errors on claude.ai, Claude Code, Claude Cowork and the Claude API. The errors that began at 14:00 UTC were mitigated at 14:36 UTC, and existing conversations were mostly working again. The page describes a second issue that prevented many users from signing in (single sign-on and Sign in with Apple were unavailable). The same issue affected starting new chats, voice conversations, Claude Code and Cowork sessions, purchases and file uploads. It asked users who were already signed in not to sign out. The page now lists the incident as resolved. The excerpt available to us does not give the time of final resolution.

Sources: Claude API release notes, September 30, 2026: Sonnet 4.5 deprecation — Anthropic (release notes) · Claude — Elevated errors on claude.ai, Claude Code, Claude Cowork and the Claude API — status.claude.com

Claude Code: fewer prompts to puzzle over, more control

Release 2.1.286 adds a count such as "2 of 5" to the permission prompt when several requests stack up. It also adds mouse support for the "N more" rows of lists in fullscreen mode: clicking one jumps to that end of the list, with hover and pressed states. A third change fixes several Claude Code processes and IDE extensions each opening a login browser when gcpAuthRefresh or awsAuthRefresh credentials expire.

Release 2.1.285 adds the CLAUDE_CODE_DISABLE_WEB_FETCH environment variable to turn off the WebFetch tool. It also adds claude --desktop, which opens the Claude desktop app on the current directory or, with --continue or --resume, on a session. A new claude plugin configure command shows a plugin's options and which are unset, or saves new values read from stdin with --values-stdin. Taken together, these changes concern what the agent is allowed to do and how clearly it presents what it asks.

Sources: Claude Code 2.1.286 release notes — Anthropic (Claude Code) · Claude Code 2.1.285 release notes — Anthropic (Claude Code)

Anthropic's writing on work, expectations and cyber capabilities

Anthropic published several pieces beyond product news. Two items address what work robots could do: "What work can robots do?" on its research site, and an item titled "Can we predict the jobs robots will do?". Others are titled "What do you want from AI?" and "GLM-5.3 and the spread of advanced cyber capabilities".

The items for these pieces carry no summary, only titles. This edition therefore reports no findings or conclusions from them. The titles alone show a range of subjects: the labor-market effects of robotics, what people expect from AI, and how advanced cyber capabilities spread.

Sources: What work can robots do? — Anthropic (research) · Can we predict the jobs robots will do? — Anthropic · What do you want from AI? — Anthropic · GLM-5.3 and the spread of advanced cyber capabilities — 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.