Claude for Business · Updated 13 October 2026
How will Claude affect my business?
Claude changes the reading, writing and analysis work in a company first. As of October 2026 it can take a whole task, such as a report built from your own files, and hand back a finished document, spreadsheet or deck. It reads long documents in one pass, and the cost of using it fell with every release this autumn.
Whether that turns into value depends on your people. The software reaches every desk on the same day; the habit of handing it real work reaches very few of them without help. This page is updated with every release that matters, and the newest ones are directly below.
Latest
What shipped most recently.
The newest releases from What’s New in Claude, each with what it means for a company and what to do that week.
- 7 October 2026: Claude Haiku 5.5, cheaper Sonnet 5.5, and API credits on Team plans
- Anthropic released Claude Haiku 5.5 on 7 October 2026: its cheapest and fastest model, about 75% cheaper to run than Haiku 4.5. The same day, Sonnet 5.5 cache reads were halved and Max and Team plans gained monthly API credits. What changed, what it means for a company, and what to do this week.
- 28 September 2026: Claude Sonnet 5.5, and the Sonnet 4.5 retirement date
- Anthropic released Claude Sonnet 5.5 on 28 September 2026: 30% faster than Sonnet 5, up to 30% cheaper per task, and strongest at everyday documents, slides and spreadsheets. Sonnet 4.5 retires from the Claude API on 30 November 2026. What changed, what it means for a company, and what to do this week.
- 22 September 2026: Claude Opus 5.5, the new default for serious work
- Anthropic released Claude Opus 5.5 on 22 September 2026: performance at the level of Claude Fable 5.1 on most work, 40% cheaper to run than Opus 5, clearer writing, and higher usage limits on paid plans. What changed, what it means for a company, and what to do this week.
Every release, newest first: What’s New in Claude.
Where it shows up
how will claude affect my business
Five places your company will feel it.
In the order we see it arrive inside the firms we work with.
- Whole tasks, handed off
- Since September, Claude chat and Claude Cowork are one conversation. People describe the outcome and Claude does the steps: searching, reading files, running code, assembling the result. The work that moves first is the work that used to take an afternoon of copying between tools.
- Finished documents, decks and designs
- Claude Docs, Slides and Design produce the file itself, edited by Claude and your team together, with every change attributed. Recurring documents, like the weekly report, are the first to move.
- Long documents read in one pass
- The current models take a 1M token context window, enough for a long agreement and its amendments together. The bottleneck shifts from reading to checking what was read.
- Repetitive work at a lower price
- Claude Haiku 5.5 costs about 75% less to run than Haiku 4.5, and Opus 5.5 about 40% less than Opus 5. Workflows shelved on cost are worth pricing again.
- New work for your IT team
- Model retirements, default settings that switch on by date, and permission settings that decide how much Claude does without asking. Each one needs an owner.
Which model
which claude model should my company use
Four models, and what each one is for.
Prices are Anthropic’s, per million input and output tokens on the Claude API. Anthropic’s own documentation recommends starting with Opus 5.5 for most work.
- Claude Opus 5.5
- The default for judgment: reading an agreement, forming a view, a long piece of analysis. $4 and $20.
- Claude Sonnet 5.5
- Everyday drafting and formatting: board packs, client updates, slides and spreadsheets. Faster, and cheaper per task. $2 and $10.
- Claude Haiku 5.5
- High-volume, well-defined work: classify, extract, summarise, route. From $0.10 and $0.50.
- Claude Fable 5.1
- The most demanding reasoning and long-running agentic work, when Opus 5.5 at higher effort still falls short on your own tests. $10 and $50.
Dates
The dates your IT team needs this quarter.
Each one is Anthropic’s, from the posts above.
- 15 October 2026: Claude Docs on by default for Enterprise
- Owners can switch it on earlier. Settle the rule on link sharing first.
- Rolling: chat and Cowork merge
- Pro and Max first. Enterprise admins get at least 30 days' notice before their organization changes.
- 30 November 2026: Claude Sonnet 4.5 retires from the API
- Anything calling claude-sonnet-4-5-20250929 needs a tested replacement before then.
What does not change
will ai replace jobs at my company
A person still reads it before it goes out.
Every release this autumn made Claude do more on its own. None of them changed who is accountable for what leaves the building. The firms that get the most from Claude are the ones where people know what may go in, check what comes out, and treat the output as a first pass with its sources beside it.
That is a people problem before it is a software problem, which is the work we do: workshops on your own documents, a retainer that follows your teams through each release, and the one page of rules your people actually read.
Questions
The questions behind the question.
What leaders ask us once they have read the headline, answered in full.
How will Claude affect my business?
Claude changes reading, writing and analysis work first. As of October 2026 it takes whole tasks and returns finished documents, spreadsheets and decks, reads long documents in one pass, and costs less with each release. The value depends on whether your people learn to hand it real work and check what comes back.
Will Claude replace people at my company?
What we see inside firms is that Claude changes the shape of the work before anything else: first drafts, reformatting and first-pass reading move to Claude, and people spend more of the day checking, deciding and talking to clients. How that plays out in headcount is a decision for each firm's leadership.
Which Claude model should my company use?
Opus 5.5 for work that needs judgment, Sonnet 5.5 for everyday drafting and formatting, Haiku 5.5 for high-volume well-defined tasks, and Fable 5.1 for the most demanding reasoning. Anthropic's documentation recommends starting with Opus 5.5.
How much does Claude cost for a business?
Seat plans such as Team and Enterprise are priced per person by Anthropic. Use through the Claude API is priced per token: per million input and output tokens, $4 and $20 for Opus 5.5, $2 and $10 for Sonnet 5.5, and from $0.10 and $0.50 for Haiku 5.5. Team and Max plans include a monthly API credit since 7 October 2026.
How do I keep up with Claude releases?
What's New in Claude on this site lists every release that matters to a company, newest first, each with what changed, what it means and what to do that week, sourced from Anthropic's own release notes.
Want Claude working inside your firm?
We run Claude workshops on your own documents and stay on as your teams adopt each release. Write to us and say what you are trying to get done.
Subject line: Claude for our business