Question 1

What is Meta Muse Wonder?

Accepted Answer

Muse Wonder is Meta’s frontier AI model, released on October 12, 2026. Meta calls the Muse Wonder models its most capable to date, built for personal and professional agentic tasks such as reviewing leases, handling insurance claims and building apps.

Question 2

Who makes Muse Wonder?

Accepted Answer

Meta. It builds on Muse Spark, the first model from Meta Superintelligence Labs, and was trained at Meta’s Prometheus data center in Ohio.

Question 3

How do I use Muse Wonder on Runable?

Accepted Answer

Sign in to Runable, start a new task and choose Muse Wonder in the model menu next to the prompt box. Describe what you need and Runable runs it on Muse Wonder.

Question 4

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

Accepted Answer

Muse Wonder Full is a larger version that posts higher scores in Meta’s table but is only available in Meta’s Muse Code subscriptions, not through the API. The numbers on this page are for the standard Muse Wonder.

Question 5

Muse Wonder vs Claude Opus 5.5: which is better?

Accepted Answer

They have the same API list price, $4 input and $20 output per million tokens. In Meta’s table Muse Wonder leads on DeepSWE 1.1 and SWE-Marathon, while Opus 5.5 leads on OSWorld, HealthBench Pro and Terminal-Bench. Artificial Analysis ranked Opus 5.5 higher overall but measured Muse Wonder as faster.

Question 6

Do I need a Meta account?

Accepted Answer

No. Muse Wonder runs on your Runable plan and credits, so you do not need a Meta account, a Muse subscription or an API key.

Question 7

What are Active Memory and Active Focus?

Accepted Answer

They are Meta’s names for two Muse Wonder behaviours. Active Memory carries relevant details across conversations. Active Focus keeps the model working toward a goal when a task gets hard: trying other approaches, recovering from errors and checking in before important actions.

Question 8

What is the context window of Muse Wonder?

Accepted Answer

1 million tokens. Meta has not published a maximum output length.

Model	Best for	Speed	API price in / out
Muse Wonder	Long-running agents, memory	148 tokens/s	$4 / $20
Opus 5.5	Long, complex agent work	96 tokens/s	$4 / $20
GPT-6 Astra	Highest intelligence, computer use	57 tokens/s	$10 / $50

Meta Muse Wonder

Made on Runable

Muse Wonder on Runable

Keeps going

Expert topics

No Meta account

How to use Muse Wonder on Runable

Select Muse Wonder

Describe the job

Edit and ship

One prompt, finished work

Why choose Muse Wonder

Strong coder

Fast for a frontier model

Remembers what matters

Plans and delegates

Top ten overall

Big context

Muse Wonder specs

Muse Wonder vs other flagships

The only AI agent you need

Muse Wonder FAQs

More AI models on Runable

Put Muse Wonder to work