Claude News
FR
Live

Claude, the essentials — edition of October 8, 2026

Claude Haiku 5.5 becomes the default fast model as agent guardrails tighten

By ·

Anthropic released Claude Haiku 5.5, a 1M-context model built for high-volume, latency-sensitive work. Claude Code adopted it as its default Haiku, and the same release notes tightened web access for Managed Agents.

Key points

Haiku 5.5 arrives with a migration warning attached

Anthropic's API release notes announce Claude Haiku 5.5 (claude-haiku-5-5) as its most capable model tuned for high-volume and latency-sensitive work. It has a 1M-token context window, 128k max output tokens, and adaptive thinking controlled by the effort parameter. Claude Code 2.1.293 added the model and describes it as the new default Haiku on the Anthropic API, priced at $0.10/$0.50 per million tokens, or $0.50/$2.50 for prompts over 100K.

The same notes caution that this is not a drop-in swap. Code written for Haiku 4.5 can break on Haiku 5.5: manual extended thinking via budget_tokens returns a 400 error. Adaptive thinking is on by default, so a response can begin with thinking blocks. Anthropic also says the same text counts as more tokens, so teams that budget by token volume should re-measure before moving over. The company points developers to its migration guide.

Sources: Release notes: Claude Haiku 5.5 launch and migration caveats (October 7, 2026) — Anthropic · Claude Code 2.1.293 — Claude Code (GitHub) · Introducing Claude Haiku 5.5 — Anthropic

Managed Agents get a narrower web, and the SDKs get computer-use scaffolding

For Claude Managed Agents, a cloud environment with limited networking now applies its allowed_hosts to the web_search and web_fetch tools. A web_fetch call for a host that allowed_hosts does not match returns a url_not_allowed error to the agent. Creating a session fails with a 400 error if an enabled web tool's allowed_domains contains an entry outside allowed_hosts, and so does a session update that adds one. An allowed_hosts entry matches one exact host unless it starts with *.

Separately, web_fetch now fetches only URLs that have already appeared in the session. That includes a URL in a user message, in a web_search result, or in a page web_fetch returned earlier. Anthropic's stated reason is that this reduces the risk of data exfiltration. On the developer side, the Python and TypeScript SDKs gained beta classes for the browser use tool and the computer use tool. A developer subclasses one, writes one method per tool against their own browser or desktop automation, and the SDK runs the tool loop and the URL and file policies the developer sets.

Sources: Release notes: Managed Agents web tools and SDK browser/computer use classes (October 7, 2026) — Anthropic

Claude Code tightens how hooks are interpreted

Claude Code 2.1.294 fixes a case where prompt and agent hooks written as instructions, such as "Block commands that...", allowed the very commands they were meant to block. It also improves how prompt hooks on Stop and SubagentStop written as instructions, such as "Carry on if the build is broken", are judged, so that Claude is less likely to stop early. Both changes concern hooks phrased as natural-language instructions rather than as conditions.

Version 2.1.293 added two smaller items for people extending the tool. The subagentStatusLine payload now includes agentType, so scripts can tell custom subagent types apart. For mods, $.tool.register gained an isDeferred option: set to false, it lists the tool's schema in the prompt from the start instead of behind tool search.

Sources: Claude Code 2.1.294 — Claude Code (GitHub) · Claude Code 2.1.293 — Claude Code (GitHub)

Costs, plans and oversight: cheaper cache reads, API credits, a wider Compliance API

Anthropic lowered the price of prompt-cache reads on Claude Sonnet 5.5 from $0.20 to $0.10 per million tokens, or 0.05x the base input price instead of 0.1x. Cache writes and all other prices are unchanged. Claude Max and Team plans now also include monthly API credits, and the release notes point to a separate page explaining how to claim them.

On the oversight side, the Compliance API chat endpoints now also return chats from the unified Claude experience. This is in beta for Claude Enterprise organizations and uses the existing Compliance Access Key. The status page shows an ongoing incident with elevated errors affecting usage data on the platform.claude.com Usage and Rate limits pages and in the Admin API usage report. A mitigation for the Usage page and the usage report is rolling out. Once it is in place, the most recent usage may take about 30 minutes to appear. Anthropic states that core functionality is not affected.

Sources: Release notes: Sonnet 5.5 cache pricing and Max/Team API credits (October 7, 2026) — Anthropic · Release notes: Compliance API and unified Claude chats (October 8, 2026) — Anthropic · Claude: Elevated errors on platform.claude.com — status.claude.com

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.