Meta launched Muse Wonder and Muse Wonder Full, two new models for Muse Code members, starting today. Muse Wonder is also accessible through the Meta Model API, making it suitable for teams that want to experiment with the model directly in development work.
Meta claims that both models can compete with leading models from OpenAI and Anthropic, based on internal results across several benchmarks. However, there are not yet any results from external sources to verify those claims. The next step is to use Wonder to power Muse for members, though no clear timeline has been announced.
Muse Wonder marks Meta’s next step in the AI arena
After launching Muse Spark, the first model from the restructured Superintelligence Labs team, along with the Muse AI agent, Meta is moving forward with Muse Wonder and Muse Wonder Full.
This launch reflects Meta’s push to reclaim its position as a major player in the AI market. The company is investing heavily in talent and infrastructure while expanding its lineup to include Muse Image and AI agents focused more on practical use cases.
From Muse Spark to models Meta says can compete with the top players
Muse Spark is the first model from the restructured Superintelligence Labs. Muse Wonder and Muse Wonder Full are now available to Muse Code members, while Wonder can also be accessed through the Meta Model API.
| Factor | Muse Spark | Muse Wonder | Muse Wonder Full |
|---|---|---|---|
| Status | Launched | Launched | Launched |
| Access channels | Not specified | Muse Code and Meta Model API | Muse Code |
| Meta's intended role | Superintelligence Labs' first model | Planned to power Muse | One of the new models for Muse |
Meta has not specified exactly when Wonder will begin powering Muse, but plans to start with members first.
What can you do with it in practice right now?
Muse Code members can now use Muse Wonder and Muse Wonder Full through the service. Wonder is also available through the Meta Model API, making it suitable for people who want to connect the model to their own work.
For Muse, the integration is still part of Meta’s plans. The company intends to begin with members, but has not said when it will start or which parts of Muse Wonder will power. For now, it is better viewed as model readiness rather than a fully available feature.
Performance claims still need to be treated cautiously
Meta claims that Muse Wonder can compete with leading models, based on the company’s own internal test results. It is therefore worth waiting for results from other sources before concluding which models it outperforms.
| Factor | Muse Wonder | GPT-6 Astra | Claude Opus 5.5 | Claude Fable 5.1 |
|---|---|---|---|---|
| Performance claim | Meta claims it can compete | Model cited by Meta for comparison | Model cited by Meta for comparison | Model cited by Meta for comparison |
| Source of test results | Meta's internal test results | Used as a reference model group | Used as a reference model group | Used as a reference model group |
Gemini 4 Argon is not yet clearly comparable because the model’s capabilities are still being revealed.
Strengths and limitations evident from the launch information
Pros
- +Muse Code members can use Muse Wonder and Muse Wonder Full
- +Muse Wonder is accessible through the Meta Model API
Cons
- −Meta has not disclosed details of its internal test results
- −Its specific capabilities and rollout timeline for Muse have not been announced
Costs the original announcement has not yet clarified
The announcement does not specify the cost of Muse Code or the Meta Model API, including any additional expenses, so the total cost cannot yet be determined.
To assess whether it offers good value, it is best to wait for Meta to provide pricing details and usage terms.
How much will Wonder change Muse?
Meta plans to initially use Wonder to power Muse for members. Wonder is already available to Muse Code members and can also be accessed through the Meta Model API.
In the next phase, Meta will develop specialized models to expand access to all Muse users. For now, however, there are no details on what additional capabilities Muse will offer or when the rollout will begin, so its real-world impact cannot yet be assessed.