Muse Wonder Full: price, access and usage limits
Muse Wonder Full is available through a Muse Code subscription. Meta lists access on Power Usage at $50 per month and Maximum Usage at $200 per month. Full does not have a public pay-as-you-go Model API rate.
Published USD prices checked October 14, 2026. Verify the price, tax and benefits offered to your account before subscribing. This guide does not report a paid hands-on test.
The plans that include Full
The official Muse Wonder page lists these two subscription offers. Usage multipliers are relative to Everyday Usage.
| Plan | Monthly price | Advertised usage |
|---|---|---|
| Power Usage | $50 | 20× Everyday Usage |
| Maximum Usage | $200 | 80× Everyday Usage |
Maximum costs four times as much as Power and advertises four times its relative usage. That arithmetic does not tell you how many of your tasks fit into either plan. Workload size and the plan's actual usage accounting still matter. In particular, neither multiplier is an unlimited token allowance.
Before spending $200, identify the limit that would prevent you from completing work on Power. If you have not measured one, start your evaluation with the lower-cost option that meets your requirements. Reassess after recording real task completions and interruptions.
Full is a separate model choice
Meta's availability documentation distinguishes ordinary Wonder from Full. Ordinary Wonder is available on Model API and throughout the Muse Code subscription tiers. Full requires Power or Maximum. Setting ordinary Wonder's reasoning effort to max does not select Full.
When following a tutorial, check both the selected model and the way the session is authenticated. A screenshot that says "max" is not enough to establish Full access. Likewise, a successful ordinary Wonder API call does not verify a Full subscription.
Muse Code is Meta's coding agent for the terminal and CI. It can plan changes, edit project files and run commands. Its documentation describes workspace trust, approvals and an operating-system sandbox. Start any evaluation in a repository where you can inspect changes and undo them.
For a service you are building around direct API calls, use the ordinary Muse Wonder API budget. Do not base that service's economics on a Full subscription or assume the subscription credential can be reused in unrelated applications.
A practical first-month evaluation
Use tasks you can judge. A bug fix with a reproducer is easier to evaluate than "improve the app." Give the agent a failing case and an expected outcome. After it finishes, review the diff, run the relevant tests and record any manual repair. A passing build alone may miss the original bug.
Keep a short worksheet with the task, selected model, acceptance result, review minutes and any limit interruption. Use the same acceptance standard for work completed with your existing tools. Provider benchmark claims can help choose what to investigate, but they cannot supply these measurements for your repository.
| Accepted tasks in a month | $50 Power plan | $200 Maximum plan |
|---|---|---|
| 20 | $2.50 per task | $10.00 per task |
| 100 | $0.50 per task | $2.00 per task |
These are divisions of the subscription price, not promised throughput or measured performance. Include the time you spend checking results when deciding whether the plan earns its place in your workflow. A rejected change still consumed some of your time and allowance.
Subscription credentials and usage limits
The subscription rules tie the plan to the credential connected during Muse Code onboarding. Additional API keys use pay-as-you-go billing. Benefits and availability can vary by region; the onboarding offer is the relevant check for your account.
Reaching a plan limit means upgrading or waiting for the allowance to refresh. On Maximum, Meta says you must wait. If your work has a deadline, plan around that possibility instead of treating the monthly fee as a guarantee of continuous execution.
Authentication can also explain an unexpected bill. Meta's authentication guide gives an environment API key priority over a stored key, and a stored key priority over browser sign-in. Check which credential the session actually uses before attributing a charge to the subscription. Keep key values private.
Review Contributor mode before sending project files
Muse Code subscriptions default to Contributor mode, which exchanges permission to use content for model training for more usage. You can turn it off in settings; Meta says that choice applies to future content. Both Wonder variants support either mode on their eligible plans.
Make this choice before an evaluation involving client code or internal documents. Confirm your team's policy and the account setting, then decide what material to include. This guide does not treat a subscription purchase as permission to upload someone else's confidential work.
Before checkout and before renewal
Confirm Full appears in the plan offered to your account, review the actual checkout total, and check the content-use setting. Then choose a review date before renewal. Define what would justify another month: fewer blocked tasks, acceptable review time or enough completed work to cover the fee.
Manage cancellation in Meta Accounts Center at least 24 hours before renewal, per Meta's subscription guide. Access continues until the billing cycle ends.
AI API Prices reviewed the official model, subscription and authentication documentation on October 14, 2026. For token-priced application work, the Muse Wonder price guide has separate cache and web-search examples.