Claude Duet's Release Date: What Anthropic Has Dated, Sequenced, and Left Open

The fastest-rising query attached to Claude Duet is the one Anthropic has answered least directly. People want a release date. What the company has actually published, on the Claude Duet product page, is narrower and more specific: a research preview on a waitlist for Claude Pro and Max subscribers in the United States, running on Claude Opus 5.5, with Team and Enterprise plans listed as "not included yet". There is no launch calendar inside that sentence, and this article does not invent one. The more useful exercise is to separate what carries a date, what carries only an order, and what carries neither.

The distinction is not pedantry. A release date is a promise about the future that a company can be held to. A waitlist is a gate with conditions attached. A stage label like research preview is a description of where a product sits today. Anthropic has given us the second and third and none of the first, and most of the search interest around the product is really a question about which of the three it is dealing with.

The only date attached to the product

The announcement article went out on 21 October. That is the last part of the Claude Duet story with a day attached to it. The product page is a separate document, and it is where the operational detail lives, including a FAQ that the announcement does not carry. When the page names the audience, the plans, the platforms or the limits, those statements are current as of the day the page was read, not as of a schedule.

"Research preview" is a stage. "Waitlist" is a gate. Neither is a date, and neither implies a deadline. A team planning around Claude Duet cannot mark a calendar from either word. What a team can do is read the eligibility sentence precisely, because it is written with unusual care.

What the waitlist actually gates

The page's call to action is "Join the waitlist", and its FAQ states the eligibility in one sentence: "Claude Duet is a research preview on a waitlist for Claude Pro and Max subscribers in the US." Three conditions are stacked there. You need a qualifying plan, you need to be in the United States, and you need to be let through a waitlist that Anthropic does not describe. The same answer adds the sentence that matters more than any rumour: "Team and Enterprise plans are not included yet."

That "yet" is doing a lot of work, and it is worth being clear about what it does and does not say. It says the exclusion is not intended to be permanent. It does not say when the exclusion ends, whether it ends for both plans at once, or what the order of the rollout will be. Three words of forward-looking language are the entire published schedule for the business tiers.

Duet also has no separate price during the preview. It is included in the plan you already hold rather than sold alongside it, which means the only prices that matter are the plan prices. Anthropic lists Pro at $17 a month billed annually, or $200 up front for the year, or $20 month to month. Max comes in two tiers, 5x at $100 a month and 20x at $200 a month. Those are plan prices, not Duet prices. Nobody can buy Duet on its own today, so a "Duet price" does not exist to compare against anything.

The things that carry no date at all

Four questions follow naturally from the eligibility sentence, and for each one the honest answer is the same word.

• A release date outside the United States — not published. The page defines the preview by country and says nothing about a second country.

• Any word on an EU or UK date — not published in either direction. The page neither promises one nor rules one out, which is a different and less satisfying answer than "no".

• A waitlist position or queue length — not published. There is no number a subscriber can check, no place in line to watch move, and no stated number of people who have been invited.

• A Duet-specific usage allowance — not published. The usage FAQ says the preview shares your plan's limits with the rest of Claude, and that background checks count toward those same limits. There is no separate quota to plan around.

Each of those is a finding rather than a gap in the reporting. Anthropic has published a product page detailed enough to explain how payroll review works and careful enough to say that larger tasks require more model turns, and it has still not published a single date for anything past the current stage. That is a deliberate level of specificity, not an oversight.

What the search data shows, and what it does not

Google Trends, worldwide, past seven days, read on 23 October 2026, shows how widely that distinction is being missed. The seed "claude duet" returns 19 rising queries, and every one of them is marked Breakout, the label Trends applies to its fastest-rising terms. The list includes "claude duet release", "claude duet uk" and "claude duet sign up" — three phrasings of the same impatience, arriving in the same week.

The regional split is starker than the query list. On the Trends index, China leads at 100, followed by Israel at 70, Singapore at 55, the United Arab Emirates at 53 and the United States at 38, across 66 regions with low-volume regions included. The country where the preview actually runs ranks fifth in search interest for it. A separate seed, "claude", returns "claude duet" as Breakout and "claude duet app" at plus 2,050 per cent, which is the same signal in a different table.

One caution belongs here because it changes how the numbers should be read. Trends also surfaces a rising topic called "Claude Paul Duet", a person whose name collides with the product query. Some of that volume is not about the product at all. The regional figures should therefore be described as search interest in the query, not as demand for the app, and any argument that depends on treating them as adoption evidence is weaker than it looks.

The shape survives the caution, though. A product confined to one country and one stage is being searched for in dozens of countries and phrased three different ways, which tells you the interest is running ahead of the rollout rather than following it.

Why "release date" is the wrong shape of question

Three different questions are hiding inside those two words, and Anthropic answers each one differently.

The first is "when can I use it". The published answer is a waitlist, which is not a date but a process with conditions. The second is "when does it stop being a research preview". The published answer is nothing at all; no exit date appears anywhere on the page. The third is "when does it reach my country or my plan". The published answer is "not included yet" for business plans and a country name for everyone else, with no timeline attached to either.

A reader who wants a date is really asking Anthropic to convert one of those answers into a commitment. The company has chosen not to, and it has been consistent about it: the page states the stage, the audience and the limits, and stops. Reporting a date would require inventing one, so the accurate answer to a search for "claude duet release" is that the release has not been scheduled in public.

What the page does tell you you would get

None of that means the page is thin. It is unusually concrete about the work Claude Duet is meant to take over, which is useful for anyone deciding whether to care about the waitlist at all.

The morning brief is the centrepiece. Duet briefs you every morning, drafts next steps, and then goes quiet until something needs you. It flags what needs attention in your inbox and calendar and drafts replies for you to approve. It finds unpaid invoices and drafts the reminders, prepares payroll for you to review, and gets contracts ready to send. It can put together a brief, a deck or a dashboard. It reaches out when a task is done or needs your approval, and it works with more than 6,000 connectors and plugins.

The surface is a standalone app for iOS and Android, plus the web at claude.ai on a computer. The limits are stated with equal care. During the preview, Duet shares your plan's usage limits with the rest of Claude and uses them differently from a typical conversation; background checks count toward those limits, and larger tasks such as building a financial model or a deck require more model turns and more background work. The page is also candid about failure modes: it may misread a thread or miscalibrate what is urgent, and it can miss something that arrives between checking in with you.

Read together, those paragraphs describe a product that is being tested in the open rather than launched. A launch date would be the least interesting fact about it. The interesting facts are the ones already written down: what it does, what it costs you in plan usage, and where it is known to fail.

The part you can have without the waitlist

Claude Duet is powered by Claude Opus 5.5, and the model is a separate thing from the app. OrcaRouter hosts the model. We do not host Claude Duet, we do not run its waitlist, and no amount of routing gets you the Duet app — that boundary is worth stating plainly, because the two are easy to confuse when one is the headline on the other's page.

What we do host is Claude Opus 5.5 itself, at $4.00 per million input tokens and $20.00 per million output tokens, with a 1M-token context window and cache reads at $0.20 per million. It sits behind one OpenAI-compatible endpoint alongside more than 200 models, on one key, with automatic failover across providers. We pass provider list prices through at zero markup, so a vendor price change is live on our side the same day.

That is not the same product. An API call gives you the model that reasons through a task; it does not give you the morning brief, the inbox monitoring, the approval step or the 6,000 connectors. Anyone who needs Duet's workflow still needs Duet. Anyone who needs the reasoning behind it, and already has their own workflow, does not need to wait for a country list to change.

How to tell a schedule from a wish

A date has a day attached. A sequence has an order attached. A waitlist has neither, and most of the writing about Claude Duet's timing has been an attempt to convert the third into the first.

The strongest forward-looking statement on the page is still "Team and Enterprise plans are not included yet", and it is not a date. The strongest geographic statement is "in the US", and it is not a date either. If Anthropic publishes a schedule, it will have to change one of those sentences to do it, which makes them the lines worth watching rather than any projection built on top of them.

Until one of those sentences changes, the answer to "when does Claude Duet come out" is that it already has, in the only form Anthropic has offered: a research preview, on a waitlist, in one country, on two plan tiers, with everything beyond that listed as not published. The page is not hiding a date. Read the three statements together — research preview, waitlist, not included yet — and they say the same thing at three levels of specificity: nothing past the current stage has been scheduled in public.