Engage Era Work with us
Meta's New AI Model Costs Less Than Half What Its Top Rivals Charge
Image: Meta
AI

Meta's New AI Model Costs Less Than Half What Its Top Rivals Charge

Meta's first frontier model scores on par with Claude Fable 5.1 and GPT-6 Astra by Meta's own testing, at $4 per million input tokens instead of $10. It will soon power the Muse app.

The short versionOn October 12, 2026, Meta launched Muse Wonder and Muse Wonder Full, which it calls its most capable models. Muse Wonder costs $4 per million input tokens and $20 per million output tokens on Muse Code and the Meta Model API; Wonder Full is only available to Muse Code subscribers. According to Meta's model card, as reported by TechCrunch, both score comparably to Claude Fable 5.1 and GPT-6 Astra, which list $10 per million input tokens. Muse Wonder will soon power the Muse app, subscribers first.

Meta just released its first frontier AI model, and it's charging developers less than half what the models it compares itself to cost. Muse Wonder and Muse Wonder Full launched on October 12. Meta calls them "our most capable models."

What happened

Meta's Superintelligence Lab released the pair on Monday. TechCrunch reports that, according to Meta's testing in its model card, "Wonder and Wonder Full achieve comparable benchmark scores to Fable 5.1 and GPT-6 Astra." Meta says Muse Wonder "will soon power experiences in Muse, and will roll out first to subscribers."

The details

Price. Meta's developer blog says Muse Wonder launches on Muse Code and the Meta Model API "at $4 input / $20 output per million tokens (USD)." Muse Wonder Full is "available exclusively by subscription in Muse Code." Anthropic and OpenAI list Claude Fable 5.1 and GPT-6 Astra at $10 per million input tokens. It isn't the cheapest model out there: TechCrunch says Wonder is roughly equivalent in cost to GPT-6 Sol or Opus 5.5.

Memory. Meta is leaning on what it calls Active Memory: "carrying important context about you across conversations."

Where it was built. Per TechCrunch, the models were trained at Meta's Prometheus data center in central Ohio, which came online earlier this year.

Mark Zuckerberg framed the bigger picture on the Cheeky Pint podcast with John Collison: "Because it's not just the people anymore. Like, now agents are effectively going to have accounts."

Why brands should care

Meta introduced Muse, its personal AI agent, on September 8. A stronger model behind it means more people asking Muse to find a plumber, book a table or compare prices, and Muse answering with whatever it can find about your business.

What to do

Our suggestions, not Meta guidance:

  1. Ask Muse to find your business. Check that it gets your hours, prices and booking link right.
  2. Fix the sources. Make sure your Facebook Page, Instagram profile and website say the same thing.
  3. Watch the price war. If you pay for AI tools, cheaper models tend to show up in them over time.

Sources

FAQ

How much does Meta Muse Wonder cost?

Meta charges $4 per million input tokens and $20 per million output tokens for Muse Wonder on Muse Code and the Meta Model API. Muse Wonder Full is only available through Muse Code subscription plans.

Is Muse Wonder as good as Claude Fable 5.1 or GPT-6 Astra?

According to Meta's own testing in its model card, as reported by TechCrunch, Muse Wonder and Wonder Full achieve comparable benchmark scores to Fable 5.1 and GPT-6 Astra.

Is Muse Wonder the cheapest AI model?

No. It costs less than half of Fable 5.1 and GPT-6 Astra per input token, but TechCrunch says it's roughly equivalent in cost to some other models, and Google's Gemini 4 Argon launched a day later at an introductory $2 per million input tokens.

Ben Meller is the founder of Engage Era and Engage Era Studio. He works with founders, artists and brands on social, content and growth. More about Ben →

Want this done for you?

Engage Era Studio builds what we cover.

See the Studio →
The Brief · Launching soon

Marketing news.
Before it's news.

The platform updates, campaigns and founder moves that matter, in a 3-minute read. The Brief is launching soon. Join the list for early access.

Free. We'll email you when it launches.