GoKawiilMeta unveils Muse Wonder and Muse Wonder Full AI models
Meta has released two new AI models, Muse Wonder and Muse Wonder Full, trained at its gigawatt-scale Prometheus data center. The company says the models build on earlier work shown with Muse Spark and post state-of-the-art results on benchmarks covering agents, memory, and coding, including domain-specific tests in law, finance and health such as Harvey's Legal Agent Benchmark.
GoKawiil's interpretation of the reporting above, not reported fact.
Meta frames the release as part of a 14-month effort to rebuild its AI stack and push toward what it calls 'personal superintelligence' for its billions of users, suggesting an ambition to embed increasingly capable agents directly into everyday consumer products. The emphasis on memory and domain expertise indicates Meta may be targeting practical, persistent assistants rather than just benchmark performance, though independent verification of these claims is not yet available.
- Meta released Muse Wonder and Muse Wonder Full, trained at its Prometheus data center
- The models reportedly achieve top results on benchmarks for agents, memory, and coding
- Meta positions the release as a step toward 'personal superintelligence' across its products
Source: research.meta.ai — Meta Ai Research, 2026-10-12
Published there as: “Muse Wonder”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.