Meta Releases Frontier Models Muse Wonder and Muse Wonder Full
Meta's Superintelligence Lab has launched two new state-of-the-art models—Muse Wonder and Muse Wonder Full—as the company’s first frontier model releases. These larger, more capable models outperform previous offerings like Muse Spark in complex tasks such as coding and achieve benchmark scores comparable to Fable 5.1 and GPT-6 Astra. The pricing for accessing these models via Meta's Model API is $4 per million input tokens and $20 per million output tokens, aligning with the cost of competing large language models like GPT-6 Sol or Opus 5.5. Muse Wonder Full will be exclusive to Muse Code subscribers and not available through the API. Meta emphasized these models' ability to retain user information over time—referred to as 'active memory.' The company's testing shows that Muse Wonder scores higher on MemoryBench than Fable, Opus, or Astra. This release marks a significant step in Meta’s AI development strategy since restructuring its efforts under Scale AI founder Alexandr Wang last year.
Tickers: $META
Topics: active memory AI models benchmarking Meta AI Muse Wonder
Read the full article at TechCrunch
Part of the story Google Launches Gemini 4 Argon, Claims Competitive Edge in AI Race (6 articles)