Muse Wonder Pricing Explained: Free vs Power vs Maximum

Muse Wonder pricing depends on where you use it. In the free Muse app, you get Muse Spark 1.3 at no cost with a weekly usage allowance. Paid plans (Power at $20/month and Maximum at $100/month) unlock Muse Wonder with much higher limits. Developers pay per token through the Meta Model API ($4 per million input, $20 per million output), and the strongest version, Wonder Full, is exclusive to Muse Code subscribers at $50 or $200 per month.

Key takeaways

• Free tier: $0. The Muse app is free and runs Muse Spark 1.3 with a weekly usage allowance (roughly 1 million input tokens). You don’t need a credit card to start.
• Power ($20/month) is the cheapest way into Muse Wonder inside the app. It unlocks the Wonder model plus a much larger weekly allowance (500 million input tokens per week, per published breakdowns).
• Maximum ($100/month) unlocks the same Wonder model with a far larger allowance (3 billion input tokens per week, per published breakdowns). Meta has not published any feature difference between Power and Maximum beyond usage.
• Muse Wonder Full, the strongest version, is available only through Muse Code subscriptions: $50/month (Power Usage) or $200/month (Maximum Usage).
• Developers get standard Wonder through the Meta Model API at $4 per million input tokens and $20 per million output tokens, with cached input at $0.20 per million.
• Meta’s CEO says the company will “work to bring this level of intelligence to everyone over time.” The free tier may eventually get Wonder, but there is no timeline.

The short version

Muse Wonder comes in three separate doors, each with its own pricing: the consumer Muse app, the Meta Model API, and Muse Code. Only some are open to everyone today. Meta announced the Wonder family on October 12, 2026, positioning it as its first frontier-class AI model, and called the two headline features Active Memory (carrying context across conversations) and Active Focus (pushing through obstacles instead of stalling).

Full pricing table

Plan / access point	Price	What you get
Muse Free	$0	Muse Spark 1.3; weekly usage allowance (~1M input tokens)
Muse Power	$20/month	Muse Wonder (rolling out first to paid plans); much higher weekly allowance (~500M input tokens)
Muse Maximum	$100/month	Same Muse Wonder access as Power; highest allowance (~3B input tokens/week)
Meta Model API (Wonder)	Pay per token	muse-wonder-1.0: $4 / 1M input, $20 / 1M output, $0.20 / 1M cached input
Meta Model API (Spark 1.3)	Pay per token	$1.25 / 1M input, $4.25 / 1M output; contributor tier $0.10 / $0.20 in exchange for training use
Muse Code (Power Usage)	$50/month	Muse Wonder access in the coding tool
Muse Code (Maximum Usage)	$200/month	Muse Wonder + Muse Wonder Full, the strongest version

Weekly allowance figures for the consumer plans come from published third-party breakdowns of the tiers; Meta itself describes the free plan only as “available for free with a usage limit,” so treat those numbers as approximate. Note that getting access involves a few extra steps depending on which door you pick.

Consumer plans: Free, Power, and Maximum

Free ($0)

The Muse app is free to download and use. On the free tier you get Meta’s Muse Spark 1.3 model with a weekly usage allowance. Meta CEO Mark Zuckerberg has cited roughly 1 million input tokens per week. When you hit the limit, you can wait for the weekly refresh or upgrade. Per published coverage, every tier includes all Muse features; the paid plans buy you more usage. Wonder is not currently available on the free tier, though Zuckerberg has said Meta will “work to bring this level of intelligence to everyone over time.” No timeline has been given.

The free Muse app hit 5 million downloads in just 22 days, per Sensor Tower data, which is exactly why Meta needs paid tiers to cover its compute costs.

Power ($20/month)

Power is the entry paid tier and, for most people, the meaningful one: it unlocks Muse Wonder inside the app. Wonder is rolling out in the app “soon,” with paid subscribers first. You also get a far higher weekly allowance (roughly 500 million input tokens per week, per published breakdowns). If you want the frontier model without writing code or paying per token, this is the door.

Maximum ($100/month)

Maximum unlocks the same Muse Wonder access as Power, with a much larger weekly allowance (roughly 3 billion input tokens per week, per published breakdowns). Here is the honest caveat: Meta has not published any feature difference between Power and Maximum. The price jump buys you ceiling, not capabilities. Unless you run very heavy, sustained agent workflows (long-running multi-app errands, large document processing, constant use), Maximum is hard to justify for one person.

Developer pricing: the Meta Model API

If you build software rather than chat in the app, Wonder is available today through the Meta Model API as muse-wonder-1.0:

• $4 per million input tokens
• $20 per million output tokens
• $0.20 per million cached input tokens
• $2.50 per 1,000 web search queries through the API

For comparison, the API rate for Muse Spark 1.3 is $1.25 per million input and $4.25 per million output. Wonder costs roughly 3.2× more on input and 4.7× more on output, reflecting that it is a much larger model. Those per-token rates sit right alongside Anthropic’s Claude Opus 5.5 ($4/$20), and well below GPT-6 Astra and Claude Fable 5.1 (both $10/$50). Alexandr Wang, Meta’s AI chief, told CNBC the API version costs more than the standard plans because Wonder is “a much larger model and at frontier level,” but would be priced “meaningfully below” leading rivals.

Spark 1.3 also has a cheaper “contributor” API tier ($0.10 input / $0.20 output per million tokens) in exchange for letting Meta train future models on your prompts and completions. There is no equivalent discounted Wonder tier at launch.

Muse Code: where Wonder Full lives

The strongest version of the model, Muse Wonder Full, is available only through Muse Code, Meta’s coding tool:

• Power Usage: $50/month: Wonder access in Muse Code
• Maximum Usage: $200/month: Wonder plus Wonder Full

According to launch coverage, Wonder Full leads the pair on coding benchmarks (including 79.0 on DeepSWE 1.1) and scores 56 on the Artificial Analysis Intelligence Index versus 52 for standard Wonder. If your interest is specifically the top-end coding model, this is the only door that opens, but you pay five to ten times the consumer subscription to walk through it.

Which should you choose?

• Try it free first. The free tier is genuinely generous, and Alexandr Wang himself has said most people should never need to leave it. You get the same features; only the allowance is smaller.
• Power ($20) if you want Wonder inside the app and you hit the free limits regularly. It is the clear value pick for the new model.
• Maximum ($100) only if you consistently exhaust Power’s allowance. You are paying for headroom, not new features.
• API if you are a developer embedding Wonder in your own product and want exact per-token cost control.
• Muse Code only if you specifically want Wonder Full for coding work and $50–$200/month makes sense against your tooling budget.

And if you are still deciding between the two models themselves rather than the prices, our Muse vs Muse Wonder comparison walks through what actually changes between Spark and Wonder.

FAQ

Is Muse Wonder free? Not today. Wonder requires a paid Muse plan ($20 Power or $100 Maximum), pay-per-token API access, or a Muse Code subscription. The free tier runs Muse Spark 1.3. Zuckerberg has said Meta will “work to bring this level of intelligence to everyone over time,” but no timeline exists.

What is the difference between Power ($20) and Maximum ($100)? As far as Meta has published: usage allowance only. Both unlock Muse Wonder in the app (rolling out first to paid plans). Power carries a weekly allowance of roughly 500 million input tokens; Maximum raises that to roughly 3 billion. No feature differences have been announced.

How do I get Muse Wonder Full? Only through a Muse Code subscription — $50/month (Power Usage) or $200/month (Maximum Usage). It is not available in the consumer Muse app or through the Meta Model API.

How much does Muse Wonder cost for developers? On the Meta Model API, muse-wonder-1.0 costs $4 per million input tokens, $20 per million output tokens, and $0.20 per million cached input tokens. Web search through the API is $2.50 per 1,000 queries.

Is Wonder cheaper than GPT or Claude? Per-token, yes in some matchups. Wonder’s $4/$20 matches Claude Opus 5.5 and undercuts GPT-6 Astra and Claude Fable 5.1 (both $10/$50). Compared with Meta’s own Spark 1.3 ($1.25/$4.25), Wonder costs several times more.

Does the free Muse tier have usage limits? Yes. Meta describes it as “available for free with a usage limit”; Zuckerberg has cited roughly 1 million input tokens per week. When you hit it, you wait for the weekly refresh or upgrade.

Can I pay per use instead of subscribing? In the app, no: it is subscriptions or the free tier. Per-use-style billing exists only through the Meta Model API, where you pay per token with no subscription.

Pricing figures come from Meta’s October 12, 2026 Wonder announcement and launch coverage from CNBC, VentureBeat, TechCrunch, and others, checked October 12, 2026. Plan limits marked “roughly” are drawn from published third-party breakdowns rather than Meta’s own pricing page, and are subject to change.