Muse Wonder
Meta · Released Oct 2026
Conditional
Muse Wonder is Meta's most capable model on its API, but it is still a preview and we could not read a single benchmark score behind the launch claims. Its list price is 3.2x Muse Spark 1.3 on input and 4.7x on output, and Meta says the full safety report is still to come. It suits teams already on Muse Spark who want to test it on long-horizon agent work, not buyers who need independently checked results.
Is it right for you?
Good for
- Long-horizon software engineering and agentic work: Meta's model docs describe it as built for complex software engineering and long-horizon agentic work (a vendor description, not a measured result)
- Mixed-media input in one request: text, image, video, audio and PDF in, text out, with a 1,048,576-token context window
- Agent plumbing on the API: Meta lists tool calling, structured output and search grounding alongside chat completion and image, video and audio understanding
- Tuning cost against effort: reasoning effort runs from minimal to max
- Predictable long-context billing: Meta's pricing page says the rate is the same whether the context window is mostly empty or almost full
Not good for
- Buyers who need independent benchmark evidence: the launch cites Harvey's Legal Agent Benchmark, FocusBench and SWE-Marathon, and all results are Meta's own (the scores were not readable in the pages we fetched)
- Production commitments today: Meta's docs say Muse Wonder is in preview, and the launch post says a full report with additional safety evaluations will be available soon
- Cheap bulk calls: Meta's pricing page shows Contributor mode as not available for Muse Wonder on the Model API, so there is no discounted data-sharing tier on this model
- Reaching Muse Wonder Full through the API: Meta says Full is available only in Muse Code with a subscription, and setting reasoning to max does not select it
Pricing
Input
$4.00 / 1M tokens
Output
$20.00 / 1M tokens
Context
1M ctx; preview; cached $0.20
No verdict changes yet
The clock starts day one. Changes land here as our verdict evolves.
Sources
Verification log
- Pricing· No changes
Automated agent