Meta One rates as of the September 16, 2026 launch Muse Video · pricing and generation limits

The prices are published. The limits are not.

You can look up what Meta One costs to the cent. What nobody outside Meta can tell you is how many clips that buys, because no quota for Muse Video has ever been published. So this page does the half that is knowable properly, and gives you a way to work out the other half for your own account.

Two panels: a published price list with four tiers, and an unpublished limits list where every figure is a question mark TWO HALVES OF ONE QUESTION, ONLY ONE OF THEM ANSWERABLE Published: the price $0 $2.99 $7.99 $19.99 Free Standalone Core Premium creator & business to $499 Not published: the limit clips per day ? reset window ? cost of a longer clip ? do discards count ? limits by country ? what Core buys in clips ? You are buying headroom of an unstated size
US rates from the Meta One launch. The quota side has never been filled in by anyone but the product itself.
Start with free

What the free tier actually covers

The app costs nothing and there is no trial, so “free” here means genuinely free rather than time-limited. What varies is how far each feature gets before it meets a ceiling.

The Mac app itself

Free, no trial, no licence, around 16 MB. Everything on this page is about compute, not software. Where to get it.

No cost

Chat, drafting, summarising

The things most people use all day sit comfortably inside the free tier. Very long sessions are the exception worth knowing about. The app-wide picture.

Rarely an issue

Image generation and editing

Metered, but generous enough that most people iterate freely. This is why settling a look in stills before generating video is good economics as well as good craft.

Metered

Video generation

Included on the free tier and the heaviest thing in the app by a distance. If you meet a limit anywhere, it will almost certainly be here first. Why it costs so much more.

First to hit a wall

Nothing that was free became paid in September

When Meta One consolidated every paid tier on September 16, 2026 gathering more than 15 million subscriptions and trial sign-ups in its first days no existing free feature moved behind the paywall. What changed is that the compute-heavy AI features got firmer rate limits on the free tier.

That distinction matters when you are deciding whether to pay. You are not buying back something that was taken away; you are buying a higher ceiling on the two or three features that have one. Meta’s stated position is that the core experience across its apps and Meta AI stays free.

If you do pay

The Meta One tiers, and which one is about video

One subscription now covers everything paid across Meta’s apps, which means the video-relevant question is narrower than the price list looks: you are choosing between Core and Premium, and only if you are actually hitting limits.

US monthly rates as of the September 16, 2026 launch; prices vary by region, app and account type, and Meta has not published how much additional video usage each tier includes. Compare against Meta Verified, Instagram Plus and WhatsApp Plus if you already pay for one of those.

A subscription buys headroom, not access

No tier will make video appear where it has not rolled out. If the option is missing from your interface entirely, that is an availability gate and paying changes nothing the five conditions explains which one is closed.

Pay when you are meeting limits, not when you are missing features. The two look identical from the outside and have opposite remedies.

The real unit of spend

You pay in attempts, not in finished clips

Two things make video the expensive feature. The first is obvious: a clip is orders of magnitude more compute than an answer or a still. The second is the one that ruins budgets you generate several clips for every one you keep, and the discards cost the same as the keeper.

A finished clip is not one generation. Plan on a handful, and the arithmetic of your allowance changes completely.

This is why the craft advice and the cost advice turn out to be the same advice. Deciding the shot before you generate, asking for the shortest clip that carries the idea, changing one clause rather than rewriting, keeping every attempt so you never regenerate something you already had all of it exists to reduce the number of runs, which is the only variable you control.

It is also why “generate ten and pick the best” a perfectly good habit with images is the fastest way to spend a week’s video allowance before lunch.

One allowance drawn on by chat, images and video at very different rates, and four attempts needed to keep one clip ONE ALLOWANCE, THREE APPETITES chat an image a clip relative draw on the same pot · not a published figure AND ONE KEEPER COSTS FOUR RUNS drifted wrong motion odd audio kept all four billed Every technique that improves clips also reduces the number of runs
Proportions are illustrative, not published figures the point is the ratio, and that discards are billed too.
The unpublished half

Six limit questions, and honest answers

These are the questions people actually arrive with. Every one has an answer you can get from the product in a week of ordinary use, and none has an answer anyone can print.

The questionThe honest answerHow to find out for yourself
How many clips per day or week?No published figure, and it would vary by country, plan and rollout wave if there were.Note the day you first see a limit message and how many clips preceded it.
When does the allowance reset?Unpublished, and reset windows differ between features and tiers.Note when generation starts working again that gap is your window.
Do discarded or failed clips count?Assume yes for anything that actually generated. A run is a run.Count your own attempts against when the ceiling arrives.
Does a longer clip cost more?Not stated, but more seconds is more compute, so plan as if it does.Ask for short clips as standard good practice regardless.
How much more does Core give me?Meta says Core includes more AI usage; it has not said how much more in clips.If you subscribe, measure the new ceiling the same way you measured the old one.
Is there per-clip or API pricing?Nothing announced for video. Subscriptions are the only published route.Track the Meta Model API page rather than assuming parity.

Any page quoting you a clips-per-day number is quoting one person’s account on one day, in one country, on one tier. That is useful anecdote and not a specification see what has and has not been published.

Should you subscribe

Three signals that paying is the right call

Without published quotas, the decision has to come from your own usage. These are the patterns where a subscription reliably earns its money.

You have met the ceiling more than once, on work that mattered

Not “I ran out while playing” that is the free tier working as designed. The signal is being stopped mid-task, twice, on something with a deadline attached.

Then: Core is the tier that raises AI usage, and Premium raises it again. Start at Core and only move up if you hit that ceiling too.

Video is a recurring part of the job, not an experiment

If clips are going out weekly, the attempts arithmetic compounds: several runs per finished clip, several clips per week. That is the profile the paid tiers exist for.

Then: price it against the alternative. At $7.99 a month, Core is cheaper than almost any stock footage subscription and than most competing video tools.

You want the business features anyway

Verified badges, impersonation protection, scheduling, exportable analytics and team accounts live in the creator and business tiers from $14.99. If you were going to pay for those, the AI headroom comes along with it.

Then: check what you already pay for first Meta Verified and the other legacy plans now sit under one subscription.

And the case against: if you are generating occasionally, settling looks in stills and keeping every attempt, the free tier is genuinely usable. Meta One versus free works through the rest of the comparison.

Spending less

Four ways to need fewer runs

Since attempts are the unit, everything here is about getting to a keeper in fewer generations. All four are free and all four also improve the clips.

Settle the look in stills first

Images meter far more lightly, so iterate the palette, framing and subject there and carry the description over. The cheapest video attempt is the one you did not need. Image prompts.

Ask for the shortest clip that works

Fewer seconds is less compute and less coherence to hold, so it is likely cheaper and definitely more likely to succeed. Two wins from one habit. Length as a budget.

Change one clause, not the whole prompt

A rewrite is a fresh roll of the dice; a single-clause edit is an experiment. Experiments converge in fewer runs. The revision loop.

Save every attempt, including the near misses

Nothing is reproducible, so a deleted clip has to be re-earned with new runs and the new one will differ anyway. Storage is free; generations are not. Keeping a folder on a Mac.

FAQ

Pricing and limits questions

Is Muse Video free?

The app is free and there is no trial, and video generation is included on the free tier with usage limits. Video is the heaviest of the metered features, so it is usually the first place anyone meets a ceiling what free covers.

How many clips can I generate per day?

Meta has not published a figure, and it would differ by country, plan and rollout wave even if it had. Watch your own limit messages that is the only number that applies to you. How to measure it.

What does Meta One cost?

At the September 16, 2026 global launch, US rates were standalone plans from about $2.99 a month, Core at $7.99, Premium at $19.99, and creator and business tiers from $14.99 up to $499. Prices vary by region, app and account type the full breakdown.

Which Meta One tier do I need for video?

Core is the tier that bundles more AI usage including image generation, editing and video, and Premium raises the limits again. Neither unlocks video where it has not rolled out a subscription buys headroom, not access.

Do failed or discarded clips count against my limit?

Assume yes for anything the service actually generated. The unit of spend is attempts rather than finished clips, which is why a habit of one considered attempt beats a spray of variations the arithmetic.

Does a longer clip cost more of my allowance?

Unpublished, but more seconds is more compute, so it is prudent to assume so. Asking for the shortest clip that carries the idea is good practice on quality grounds anyway.

When does my allowance reset?

Not published, and reset windows differ between features and tiers. The practical approach is to note when a limit message appears and when generation starts working again.

Did Meta One make Meta AI paid?

No. Nothing that was free became paid when Meta One launched the compute-heavy AI features simply got firmer rate limits on the free tier, and Meta’s position is that the core experience stays free.

Is video cheaper on Mac than on the phone or web?

No. The allowance belongs to the account, not the device, and generation happens on Meta’s servers whichever surface you use nothing renders on your Mac. Switching surfaces does not buy you more.

Is a subscription worth it just for Muse Video?

Only if you are actually hitting the ceiling. The honest test is whether you have met a limit more than once in a week on work that mattered otherwise the free tier is doing its job. The three signals.

Measure your own ceiling before you pay for a bigger one

Fewer runs beats a bigger allowance.

Settle the look in stills, ask for the shortest clip that carries the idea, change one clause at a time, and keep everything. Then, if you are still hitting the wall on real work, Core is the tier to look at.

  • App free, no trial
  • Video meters hardest
  • Discards are billed
  • Headroom, not access

Independent guide not affiliated with Meta. Prices shown are US monthly rates as of the September 16, 2026 Meta One launch and vary by region; no generation quota has been published. Confirm current pricing and limits in Meta’s own documentation before subscribing.