Meta launches first frontier AI models: Muse Wonder
Original: Meta releases its first frontier model, Muse Wonder
Why This Matters
Meta enters the frontier model race directly, challenging OpenAI and Anthropic at the top tier.
Meta's Superintelligence Lab released Muse Wonder and Muse Wonder Full on Oct. 12, 2026 — its first frontier-class models, priced at $4/M input and $20/M output tokens, benchmarking alongside GPT-6 Astra and Fable 5.1.
Meta's Superintelligence Lab released two new models on Monday: Muse Wonder and Muse Wonder Full. Unlike the earlier Muse Spark, which targeted low-latency lightweight tasks, the Wonder models are larger and built for complex work including coding. Meta's internal benchmarks place them on par with OpenAI's GPT-6 Astra and Fable 5.1 across standard evaluations.
Pricing via the Meta Model API sits at $4.00 per million input tokens and $20.00 per million output tokens — roughly in line with GPT-6 Sol or Anthropic's Opus 5.5. Wonder Full, however, won't be API-accessible; it's exclusive to Muse Code subscribers. Both models will power the Muse agent.
A standout feature Meta is pushing is what it calls "active memory" — the ability to retain and synthesize user information over days, weeks, or months. Meta claims Muse Wonder now outscores Fable, Opus, and Astra on MemoryBench, a benchmark focused on continual learning and memory construction. Research scientist Kavya Kopparapu said in a launch video: "We put a tremendous amount of effort into making sure your model can remember what's important to you over days, weeks, months and beyond."
The models were trained at Meta's Prometheus data center in central Ohio, which opened earlier this year. This is the first frontier-tier release since Scale AI founder Alexandr Wang reorganized Meta's AI operations in August 2025.