Muse Wonder is a closed API model. Meta announced Muse Wonder and Muse Wonder Full on 12 October 2026. Wonder is on the Meta Model API and in Muse Code. Wonder Full is on Muse Code for subscribers. Weights, a parameter count, and a VRAM requirement are not published. An unofficial repository that is not Meta’s is not a checkpoint and is not linked here.
Price
Checked 16 October 2026 on the model page and the pricing page.
| Model id | Where | Cached input | Input | Output |
|---|---|---|---|---|
muse-wonder-1.0 |
Model API and Muse Code | $0.20 / 1M | $4.00 / 1M | $20.00 / 1M |
muse-wonder-full-1.0 |
Muse Code subscription | not listed | not listed | not listed |
Both list a 1M context on the model page. Contributor mode is not available for Muse Wonder. Standard mode does not use request content to train Meta models. There is no long-context premium. Web search grounding is $2.50 per 1,000 queries, added to the token cost. Rate limits for muse-wonder-1.0 are 3,000 requests per minute and 4,000,000 tokens per minute, per team.
Muse Code plans that include Wonder Full, from the model page:
| Plan | Price | What the page says about Wonder Full |
|---|---|---|
| Power Usage | $50.00 / month | Includes access to Muse Wonder Full; 20× the Everyday Usage plan |
| Maximum Usage | $200.00 / month | Includes the Power Usage plan; 80× the Everyday Usage plan |
Other published prices, for scale
These are separate products. They are not quality scores.
| Product | Input | Cached input | Output | Source |
|---|---|---|---|---|
| Muse Wonder standard | $4.00 | $0.20 | $20.00 | Meta pricing page |
| Muse Spark 1.3 / 1.2 / 1.1 standard | $1.25 | $0.15 | $4.25 | Meta pricing page |
| Gemini 4 Argon through 30 Nov 2026 | $2.00 | $0.10 | $10.00 | Gemini 4 Argon |
Prices are per 1M tokens. Gemini’s later rates and the free-tier limit are on that page.
Local run
The news post and the research post describe API and Muse Code access. They do not link a weight file. The research post says a full safety report will be available soon. This page does not fill that gap.
Muse Spark Open is the open-weight Muse release from 15 October 2026. Its files are large, and this site still does not run them through the LLM estimator. Models the calculator does size include Qwen3.8-27B. See the methodology for what a fit number means on this site.
FAQ
Can Muse Wonder run locally?
No public weights were published on the Meta news post, the research post, or the model page checked 16 Oct 2026. Parameter count and VRAM are not published. This site does not estimate a local fit. Muse Spark Open is a separate open-weight model.
How much does Muse Wonder cost?
Model id muse-wonder-1.0, standard mode, per 1M tokens: cached input $0.20, input $4.00, output $20.00. Contributor mode is not available. There is no long-context premium. Web search grounding is $2.50 per 1,000 queries on top of token cost. Muse Wonder Full has no per-token price. It is included with Muse Code Power at $50 per month and Maximum at $200 per month.
What is the Muse Wonder context length?
Meta's model page lists a 1M context for both Muse Wonder and Muse Wonder Full. A parameter count is not published.
When was Muse Wonder released?
Meta's news post and research post are dated 12 October 2026. The news post says Muse Wonder is available through the Meta Model API and Muse Code, and Muse Wonder Full is available for Muse Code subscribers.