Claude News
FR
Live

Claude, the essentials — edition of November 1, 2026

Fast mode for Opus 4.8 gets a removal date, while admin and analytics tooling expands

By ·

Anthropic set November 30, 2026 as the removal date for fast mode on Claude Opus 4.8 and launched a beta analytics API for usage and cost. Three Claude Code releases added gateway and permission controls aimed at organizations.

Key points

The API: a deprecation date, a new analytics beta, and larger requests

The API release notes for the last days of October mix one retirement notice with two additions. On October 30, Anthropic said it has deprecated fast mode for Claude Opus 4.8 and will remove it on November 30, 2026. After that date, requests to claude-opus-4-8 with speed "fast" will return an error. The note's stated path for anyone who wants to keep using fast mode is to migrate to Claude Opus 5.5. The notice period is about a month.

On October 29, Anthropic launched the Usage and Cost Analytics API in beta, behind the analytics-usage-cost-2026-08-24 beta header. Its /v1/organizations/analytics/usage and /v1/organizations/analytics/cost endpoints report token usage and cost for Claude Console and Claude Enterprise organizations. The same day, the existing Usage and Cost API became a legacy API. So did the usage_report, cost_report, user_usage_report and user_cost_report endpoints of the Claude Enterprise Analytics API. Anthropic says all of them remain supported and points existing integrations to a migration guide. Separately, on October 28, the maximum request size for the Messages API and the Token Counting API on the first-party Claude API went from 32 MB to 128 MB.

Sources: Release notes, October 30, 2026: fast mode deprecated for Claude Opus 4.8 — Anthropic (release notes) · Release notes, October 29, 2026: Usage and Cost Analytics API beta and legacy designations — Anthropic (release notes) · Release notes, October 28, 2026: maximum request size raised to 128 MB — Anthropic (release notes)

Claude Code: more control for administrators and for each session

Most of the entries in the three Claude Code releases concern the Claude apps gateway, the layer through which organizations route their developers. Version 2.1.312 added oidc.user_id_claim, which lets the gateway identify people by an id_token claim other than sub, such as Microsoft Entra ID's oid. Version 2.1.314 added an optional groups list to gateway upstreams, so only developers in the listed identity-provider groups are sent to that upstream, failover included. It also added guardrail_trace to Bedrock upstreams, which logs which Amazon Bedrock guardrail filter fired on a request without recording the matched text. Version 2.1.316 added pooled group budgets: one monthly budget that a group's members share, on top of each member's own limit, set with pooled_budget on the spend limits admin API.

Two further changes concern local behavior. Version 2.1.314 introduced permissions.blockWritesOutsideWorkingDirectories, under which Write, Edit and NotebookEdit refuse paths outside the working directories in every permission mode. Version 2.1.316 added two environment variables. CLAUDE_CODE_DISABLE_ULTRATHINK turns the ultrathink keyword off. CLAUDE_CODE_DISABLE_LOOPED_CROSS_CERT_DROP keeps system cross-certificates that certify each other in a loop, for servers whose private CA needs one. Version 2.1.312 also fixed a bug in which managed settings and organization policy were not fetched for the rest of the session after signing in from the VS Code extension. It added click to send as well: in fullscreen, clicking the "to send now" hint under a queued message sends it immediately.

Sources: Claude Code 2.1.316 release — Claude Code (GitHub releases) · Claude Code 2.1.314 release — Claude Code (GitHub releases) · Claude Code 2.1.312 release — Claude Code (GitHub releases)

Science partnerships and two research posts

On October 28, Anthropic announced an expansion of AI for Science. The program gives free Claude credits to researchers working on high-impact scientific projects. It began in May 2025, and the company now highlights 20 scientific foundations, institutions and research groups as partners. Anthropic says it has also launched a discounted Team plan for scientists and supported hundreds of projects across biology, medicine, physics, chemistry, climate science and mathematics. Its own examples are a shared specification developed with HHMI Janelia Research Campus, the Model Hardware Standard (MHS), now in research preview, which lets AI agents safely operate lab instruments. The other is the Allen Institute, which Anthropic says is building teams of AI agents to analyze its data, turning months of manual analysis into hours. The company describes the shift as moving from supporting individual scientists to reaching more researchers through institutions and foundations, and says many projects will also produce public goods for the wider scientific community.

Two other Anthropic publications appeared in the day's feed: "Early progress on accelerating drug development with Claude" and, on the Alignment Science blog, "Automated Subtle Emergent Misalignment Research". The material available to us for both contains only the titles. We therefore do not summarize their methods or findings, and readers should go to the original posts.

Sources: AI for Science expands to foundations and research institutions — Anthropic · Early progress on accelerating drug development with Claude — Anthropic · Automated Subtle Emergent Misalignment Research — Anthropic Alignment Science blog

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.