What is the difference between Muse Wonder and Muse Wonder Full?

Meta ships Muse Wonder (model id muse-wonder-1.0) through both the Meta Model API and Muse Code. Muse Wonder Full (muse-wonder-full-1.0) is the larger setting and is only available inside Muse Code on a paid subscription, such as the $50-a-month Power Usage plan. Muse Wonder Full is not offered through the API.

How much does the Muse Wonder API cost?

Meta lists muse-wonder-1.0 on the Meta Model API at $4.00 per million input tokens, $0.20 per million cached input tokens and $20.00 per million output tokens. That is more than three times the $1.25 / $4.25 rate of Muse Spark 1.3, so Muse Wonder is Meta's premium API tier rather than a replacement for the cheaper model.

When will Muse Wonder come to the Muse app?

Meta says Muse Wonder will soon power experiences in Muse, its personal AI agent, and will roll out to Muse subscribers first. At launch on 12 October 2026, Muse Wonder was live for developers in the Meta Model API and Muse Code, and Muse Wonder Full was live for Muse Code subscribers.

Is Muse Wonder open weights?

No. Muse Wonder and Muse Wonder Full are proprietary, closed-weight models from Meta Superintelligence Labs, reachable only through Meta's own Model API and Muse Code. Meta's open-weight Muse model remains Muse Glimmer, a 30B Apache-2.0 model released in August 2026.

How does Meta handle safety for Muse Wonder?

Meta says Muse Wonder was trained to avoid sharing private information without permission, to decline requests for real harm, and to help defend against cyberattacks, while still letting security teams find flaws in their own code. Meta releases its most capable models under its Superintelligence Scaling Framework and says a fuller safety report is coming.

Meta · 2026-10-12 · seismic

Muse Wonder — Meta's first frontier model, built for long agent tasks

Muse Wonder and Muse Wonder Full are Meta's most capable models, built for long agentic work in law, finance, health and coding. Muse Wonder is out on the Meta Model API at $4 / $20 per million tokens with a 1M-token context.

Meta AI Research announcement artwork for the Muse Wonder model series — Meta AI Research

Meta's biggest Muse models yet remember what matters to you and keep working on a hard task until it is done.

Quick facts

Maker	Meta Superintelligence Labs
Models	Muse Wonder, Muse Wonder Full
Model ids	muse-wonder-1.0, muse-wonder-full-1.0
Context window	1M tokens
Availability	Wonder: Meta Model API + Muse Code; Full: Muse Code plans only
License	Proprietary, closed weights

Pricing

Input · muse-wonder-1.0	$4.00 / 1M tokens
Cached input · muse-wonder-1.0	$0.20 / 1M tokens
Output · muse-wonder-1.0	$20.00 / 1M tokens
Muse Code Power Usage · includes Muse Wonder Full	$50 / month
Muse Code Maximum Usage	$200 / month

source ↗

What is it?

Muse Wonder and Muse Wonder Full are a new model series from Meta Superintelligence Labs, released on 12 October 2026. Meta calls them its most capable models so far and aims them at personal and professional agent tasks, such as handling an insurance claim, understanding a lease or building a working app. TechCrunch reports they are larger than Muse Spark, which was built for low latency.

How does it work?

Two trained behaviours carry the launch. Active Memory keeps context about the user across conversations and pulls back the details that matter for the current task. Active Focus keeps the model on a goal: it tries other approaches, works through errors, asks for clarification and checks in before important actions. Meta says the models were pretrained and trained with reinforcement learning at its Prometheus data center, then post-trained on long-running tasks.

Why does it matter?

For Meta, Muse Wonder is the first model it positions at the frontier rather than as a fast, cheap option, and TechCrunch notes the model card puts it in the same range as Claude Fable 5.1 and GPT-6 Astra. For developers, the API price of $4 / $20 per million tokens sits in the premium tier, so the real test is whether its memory and long-task focus save enough agent turns to pay for that.

Who is it for?

developers building long-running agents, Muse Code users

Frequently asked questions

What is the difference between Muse Wonder and Muse Wonder Full?: Meta ships Muse Wonder (model id muse-wonder-1.0) through both the Meta Model API and Muse Code. Muse Wonder Full (muse-wonder-full-1.0) is the larger setting and is only available inside Muse Code on a paid subscription, such as the $50-a-month Power Usage plan. Muse Wonder Full is not offered through the API.
How much does the Muse Wonder API cost?: Meta lists muse-wonder-1.0 on the Meta Model API at $4.00 per million input tokens, $0.20 per million cached input tokens and $20.00 per million output tokens. That is more than three times the $1.25 / $4.25 rate of Muse Spark 1.3, so Muse Wonder is Meta's premium API tier rather than a replacement for the cheaper model.
When will Muse Wonder come to the Muse app?: Meta says Muse Wonder will soon power experiences in Muse, its personal AI agent, and will roll out to Muse subscribers first. At launch on 12 October 2026, Muse Wonder was live for developers in the Meta Model API and Muse Code, and Muse Wonder Full was live for Muse Code subscribers.
Is Muse Wonder open weights?: No. Muse Wonder and Muse Wonder Full are proprietary, closed-weight models from Meta Superintelligence Labs, reachable only through Meta's own Model API and Muse Code. Meta's open-weight Muse model remains Muse Glimmer, a 30B Apache-2.0 model released in August 2026.
How does Meta handle safety for Muse Wonder?: Meta says Muse Wonder was trained to avoid sharing private information without permission, to decline requests for real harm, and to help defend against cyberattacks, while still letting security teams find flaws in their own code. Meta releases its most capable models under its Superintelligence Scaling Framework and says a fuller safety report is coming.

Try it

Model id muse-wonder-1.0 on the Meta Model API (dev.meta.ai)