Meta Launches Muse Wonder, But Independent Rankings Still Favor Anthropic’s Opus 5.5
What happened: Meta released its first frontier-class AI models, Muse Wonder and Wonder Full, on October 12.
What happened: Meta released its first frontier-class AI models, Muse Wonder and Wonder Full, on October 12. According to Meta, Wonder Full outperformed GPT-6 Astra and Claude on internal benchmarks, but independent rankings still place Anthropic’s Opus 5.5 slightly ahead. Wonder Full is available only to Muse Code subscribers, not via API, and is priced to match Opus 5.5 while undercutting GPT-6 Astra.
Why it matters: While Meta touts benchmark superiority, the lack of open API access for its flagship model limits third-party verification and widespread adoption. The move mirrors OpenAI’s strategy of gating top models, potentially slowing ecosystem growth. Independent analysis suggests the AI race remains competitive, with no clear leader across all metrics.
Source: Decrypt, TechCrunch, VentureBeat