Meta has launched two new frontier-class AI models, Muse Wonder and Muse Wonder Full, developed by its rebuilt Meta Superintelligence Labs. These models are designed for complex, long-running agentic tasks, featuring a 1M-token context window and multimodal input capabilities. While Meta claims they are on par with leading models from OpenAI and Anthropic, independent benchmarks show Muse Wonder Full tying with Claude Sonnet 5.5 on the Artificial Analysis Intelligence Index, slightly behind Claude Opus 5.5. The best-performing version, Muse Wonder Full, is not yet available through Meta's public API. AI
IMPACT Sets a new benchmark for agentic AI capabilities, potentially influencing future model development and enterprise adoption.
RANK_REASON Frontier-lab model release with system card and independent benchmark coverage. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Alexandr Wang
- Anthropic
- Claude Fable 5.1
- Claude Opus 5.5
- Claude Sonnet 5.5
- GPT-6 Astra
- Llama 4
- Meta
- Meta Superintelligence Labs
- Muse Wonder
- Muse Wonder Full
- OpenAI
- Prometheus
- Scale AI
- Zuckerberg
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →