Meta releases its first frontier model, Muse Wonder

THE NEWS

What happened

Meta has officially launched its first frontier AI models, Muse Wonder and Muse Wonder Full, marking a significant shift from its previous focus on lightweight tools to high-capability systems. These new models compete directly with top-tier industry leaders like GPT-6 Astra and Fable 5.1, offering advanced coding skills and a unique 'active memory' feature that allows the AI to retain user information over long periods. The release signals Meta's renewed commitment to developing personal superintelligence following recent leadership restructuring.

CONTEXT

Why it matters

Meta has officially entered the frontier model race with Muse Wonder and Muse Full. These large-scale models rival GPT-6 Astra in benchmark scores and introduce 'active memory' to retain user context over months. Muse Wonder is available via API, while Muse Full requires a Muse Code subscription. Trained at Meta's new Ohio data center, this marks a major strategic pivot.

AT A GLANCE

Key facts

• Meta released Muse Wonder and Muse Wonder Full as its first frontier models.
• The models achieve benchmark scores comparable to GPT-6 Astra and Fable 5.1.
• Muse Wonder costs $4 per million input tokens and $20 per million output tokens via the Meta Model API.
• Muse Wonder Full is exclusive to Muse Code subscribers and is not available through the public API.
• The models feature 'active memory' to synthesize information over days, weeks, or months.
• Muse Wonder scores higher on the MemoryBench than previous Meta models like Fable and Opus.
• The models were trained at Meta's Prometheus data center in central Ohio.
• This launch follows a restructuring of Meta's AI efforts by Scale AI founder Alexandr Wang.