A standing box for the in app rollout, updated when a check below moves:
- Meta has announced an in app date: not yet. No date appeared in the sources checked on October 27 to 30, and none appeared in the checks run for this post on October 30.
- Meta has published a rollout order or a finish date, beyond subscribers first: not yet.
- A fresh first person sighting in the October 27 to 30 record settles the current state for a typical subscriber account: not yet. The claims in that record point in opposite directions, and the section below lays them out instead of picking one.
- Next update to this box: on the next scheduled check, November 1, 2026, or sooner if Meta dates the rollout, a US upgrade screen label is sighted first hand, or a dated first person sighting from an ordinary subscriber account is published. Whichever comes first.
The short version
There is no date. That is the whole answer Meta has given, and it has not changed in the nineteen days since Wonder launched.
What is live today, on the record checked, is the developer side. Muse Wonder is available through the Meta Model API at $4 per million input tokens and $20 per million output tokens. Muse Wonder Full is not on that API. It sits with Muse Code subscribers instead. Consumer Muse tiers have not moved: free, Power at $20 a month, and Maximum at $100 a month.
What is not settled is the personal agent itself. Meta Research still frames Wonder in the future tense for Muse: it will soon power experiences in Muse, and it will roll out first to Muse subscribers. Soon and first are the only timing words Meta has put on this. No week. No month. No list of who goes in which wave.
Where the confusion starts
This site's own launch post is part of the record, so it belongs in this post too. On October 12, the launch coverage here reported a Power account running Muse Wonder by launch afternoon, and CNBC reported Wonder as already available to Muse app subscribers that week. Read that for what it was: an early wave sighting at launch, on a paid account.
The external trackers checked after that did not line up behind it. Aiagentslibrary, rechecked repeatedly from October 14 onward in this site's sweeps, still described the Muse app as running Spark, with the subscriber rollout forthcoming and no date. Dev.to, CrowdfundInsider, and others in the same sweeps said the same thing in plainer words: API and Muse Code today, app soon.
Both things can be partly true in a wave rollout. Some accounts flip early. Most do not, and the trackers describing the general state keep writing Spark because Spark is what most accounts still report. What neither side supplies is the thing a reader actually needs: a schedule. An early sighting is not a schedule, and a tracker saying soon is not one either.
Ask your own Muse which model is running it, and write down the answer with the date and your tier. That answer is a sighting for one account on one day. It is useful evidence, and this page will treat reader sightings that way below. It is not proof about your neighbour's account, and a different answer next week would not make today's answer a lie. Wave rollouts look exactly like this from inside them: uneven, undocumented, and easy to overread.
The claim that does not settle it
The newest conflicting claim arrived in the sweep late on October 29. How to snippets from All Things Muse assert that Wonder already powers Muse for subscribers, with in app access through the $20 Power and $100 Maximum plans.
Those snippets are undated. They carry no first person in app sighting of their own. Four other sources on the same results page pointed the other way: Meta Research, with its will soon power wording, Promptwire from October 12, crescitaly, and Notebookcheck. Notebookcheck said the Muse app does not have the new model yet.
A claim contradicted on its own results page does not move the watch box. It earns a paragraph here, because readers will find it in search and wonder why this site has not printed it as fact. This is why. Undated, unsighted, and contradicted is a lead to keep checking. It is not a confirmation.
The checks run for this post on October 30 found the same shape. No dated October 27 to 30 source said, with a sighting behind it, that Wonder is now confirmed live in the app for subscribers generally. The only recent in app wording was still the future tense one.
The headline that was not an arrival
One more false signal needs naming, because it will keep resurfacing in search results. On October 28, a Meta Newsroom index snippet dated October 28, 2026 carried the title Introducing Muse Wonder and Muse Wonder Full. It looked like a second launch, sixteen days after the first.
It was not one. Verification in the same sweep, three DuckDuckGo queries run between 14:12 and 14:14 CDT on October 28, found that the October 28 date appeared only in the index snippet. The standalone announcement page carried no such date. Other sources dated the same launch to October 12, which is where this site filed it as idea number 96 in the first place.
That record sits in this site's files as a separate post idea, number 117, about how a snippet is not a story. It matters here for one practical reason: if your search today shows Wonder launching on October 28, you are looking at the October 12 release wearing a new index date. Nothing in that snippet says a word about the app rollout, a tier, or a date for either.
The two labels to watch
Two labels are circulating, and neither is corroborated yet. Both are worth knowing by sight, because one of them, confirmed, would change this page.
SmartScope reported on October 26 that a Japan app upgrade screen carried a note reading Runs on Muse Wonder. That is a report of a label, from one source, in one market. No second source in the sweeps checked confirmed it, and no US screen with that wording has been sighted first hand in the record this post rests on.
Against that, musecodes.io still described the consumer app as running Spark 1.3, with Wonder listed as soon. Spark 1.3 is the model the trackers keep naming as the current state of the app. Soon is the word Meta's own side keeps using too.
So the label test is simple. If your upgrade screen, in your own account, says Runs on Muse Wonder today, that screen outranks both paragraphs above for your account. Take the screenshot, note the tier and the date, and you have the sighting this watch is waiting for. If your app still names Spark, that is also a fact, for your account, on that date.
Do not subscribe on the strength of a label report from another country, or an undated how to page. Power at $20 and Maximum at $100 buy more usage today, on the pricing record that has held flat through every sweep this month. Whether either tier flips your account to Wonder this week is exactly the fact Meta has not promised. Pay for the volume if the volume is worth it. The model flip, on current evidence, arrives on Meta's timing, not on the day you pay.
What would count as arrival
This page uses three tests, and readers can use the same three when the next headline lands.
First, a Meta date. A help page, an announcement, or an in app notice from Meta that names a date or a window for the app rollout. Blogger estimates of a window do not count, and neither does a re-dated announcement page.
Second, a first hand label. A US upgrade screen or settings screen, seen in a real account, carrying the Runs on Muse Wonder wording, with the tier and date attached.
Third, a dated first person sighting. Someone with a name or a publication behind them, saying on a stated date that their Muse, on a stated tier, is running Wonder, in terms a reader could check against their own app.
The October 12 launch week produced an early version of the third test for some subscribers. Nothing since has produced the first test at all, and nothing in the October 27 to 30 record produced a fresh sighting that generalises beyond one account. That is the gap this post will keep reporting until it closes.
What to do while you wait
Check the model in your own app, once, and note what it says. If it says Spark, you have your current state, and rechecking daily will not change the schedule Meta has not published.
Keep the free and paid decision separate from the Wonder question. The tiers are a usage decision on today's record. Wonder is a timing question with no answer yet. Mixing the two is how a reader ends up paying $20 for a rumour.
That is this week's Rollout Check, the standing reader report for this watch: have you seen a Runs on Muse Wonder label in your app? If you have, the standard above is the ask: a screenshot or a first person account, your tier, and the date you saw it. A re-dated announcement does not count, and this page will say so plainly when one arrives dressed as news. The first reader sighting that meets the standard changes the watch box at the top of this page, and the next post in this series will say what came in.
Quick answers
When does Muse Wonder reach the Muse app?
Meta has given no date. The only timing on record, from Meta itself, is subscribers first and will soon power, checked again on October 30, 2026. Developer access has been live since October 12, 2026. That is a different door.
Is Wonder already in the app for subscribers?
Unsettled, and this page will not flatten it. An early wave was reported at launch in October, including a Power account sighting in this site's October 13 launch post. Trackers checked from October 14 onward still described the app as running Spark with the rollout forthcoming. Undated how to snippets claiming subscribers already have it are contradicted on their own results page. No fresh sighting in the October 27 to 30 record settles the general state.
How will I know when my account has it?
Look for the label in your own app: an upgrade or settings screen naming Muse Wonder, of the kind SmartScope reported, uncorroborated, from a Japan screen on October 26. Or ask your Muse which model is running it and record the answer with your tier and the date. A search headline with a new date on the October 12 announcement is not that signal.
Does paying for Power or Maximum get me Wonder now?
Not on any promise Meta has published. Subscribers are said to go first, which is an order, not a date. Power costs $20 a month and Maximum costs $100 a month, and both buy more usage on the current pricing record. No source checked on October 30 promised that paying today flips an account this week.
What about Muse Wonder Full?
Full is not part of the app question on current evidence. It has no API price because it is not sold by the token. It sits with Muse Code subscribers, at the $50 and $200 Muse Code tiers reported at launch, and Meta has not announced it for the personal agent at all.
Is the October 28 Wonder headline a second launch?
No. The October 28 date appeared only in a Meta Newsroom index snippet. The launch itself is October 12, 2026. That settlement is recorded in this site's October 28 news watch.
Frequently Asked Questions
Q: Why do some sources say the app runs Spark and others say Wonder is live?
A: A wave rollout produces both answers at once, for different accounts. Undated pages also keep republishing launch week claims as if they were today's state. The sources that date their checks, Meta Research included, still use future tense for the app. The claims that state arrival as a general fact, in the record checked for this post, do not date a sighting behind it.
Q: Did Meta miss a deadline for this rollout?
A: No deadline was set, so none has been missed. That is the uncomfortable part of this story, and dressing it up would not help. Meta promised an order, subscribers first, and kept the calendar to itself.
Q: Should free users expect Wonder at all?
A: Meta has published no free tier timeline for Wonder in the sources checked. Everything confirmed or claimed about the early app waves concerns paid accounts. A free user reading a subscriber claim is reading about a different queue, on the current record.
Q: What is the difference between the API launch and the app rollout?
A: The API is a door a developer opens deliberately, priced at $4 per million input tokens and $20 per million output tokens for Wonder, live since October 12. The app rollout is Meta flipping the model underneath an existing personal agent, account by account. The first is done. The second has no date.
Q: What should readers report back?
A: One thing, in one form: a Runs on Muse Wonder label seen in your own app, or a dated first person account of your Muse naming Wonder as its model, with your tier and the date. That is the evidence that moves the watch box at the top of this page.
Sources
- Meta, Introducing Muse Wonder and Muse Wonder Full, October 12, 2026, and Meta Research model pages, framing checked October 30, 2026: will soon power experiences in Muse, roll out first to Muse subscribers.
- Muse Explained, Muse Wonder explained: Meta's first frontier model, October 13, 2026, launch week record including the Power account sighting and the API pricing at $4 and $20 per million tokens.
- This site's news watch, October 28, 2026: the October 28 Meta Newsroom index snippet settled as the October 12 release re-surfaced, on three verification queries run 14:12 to 14:14 CDT that day.
- This site's demand log sweeps, October 27 to 30, 2026: no new launch, pricing change, or in app confirmation; All Things Muse how to snippets surfaced October 29, undated and contradicted on the same results page by Meta Research, Promptwire, crescitaly, and Notebookcheck.
- SmartScope, October 26, 2026, reporting a Japan app upgrade screen carrying a Runs on Muse Wonder note, uncorroborated in the sweeps checked.
- Musecodes.io, as banked in this site's sweeps: the consumer app described as running Spark 1.3, with Wonder listed as soon.
- Notebookcheck, as surfaced October 29, 2026: the Muse app does not have the new model yet.
Comments