{"version":1,"type":"story","url":"https://digestai.news/story/meta-adds-specialized-muse-models-with-active-memory-and-active-focus","json":"https://digestai.news/story/meta-adds-specialized-muse-models-with-active-memory-and-active-focus.json","markdown":"https://digestai.news/story/meta-adds-specialized-muse-models-with-active-memory-and-active-focus.md","slug":"meta-adds-specialized-muse-models-with-active-memory-and-active-focus","headline":"Meta improves Muse and Meta AI with Active Memory and Active Focus","summary":"Meta AI Research rolled out updated versions of Muse and Meta AI on November 6, 2026, using specialized models trained with **Muse Wonder**. The current Muse model now scores **47%** on Meta’s **Personal Superintelligence Index**, up from **38%** for the previous version. These updates introduce **Active Memory** and **Active Focus**, features designed to improve context retention and task persistence in real-world scenarios.\n\n**Active Memory** reduces forgetting critical information by **2.4x**, with Meta reporting **90%** of tasks now correctly apply relevant context—up from **76%**. **Active Focus** improves task completion by **67%** when unexpected issues arise, helping agents adapt to interruptions while maintaining progress. The changes apply to both Muse and Meta AI, aiming to make them better collaborators for users navigating complex or dynamic workflows.","keyPoints":["Muse scores **47%** on Meta’s Personal Superintelligence Index, up from **38%** with the previous model","**Active Memory** cuts forgetting critical info by **2.4x**, with **90%** of tasks now using context correctly","**Active Focus** boosts task completion by **67%** when unexpected issues arise"],"whyItMatters":"Meta’s updates make its AI assistants more reliable for long-term tasks and dynamic collaboration, a key test for enterprise and productivity tools. The focus on context retention and adaptability could set benchmarks for other agent systems.","category":{"slug":"agents","name":"Agents & Tools","url":"https://digestai.news/category/agents"},"entities":{"companies":["Meta","Meta AI Research"],"models":["Muse","Muse Wonder","Muse Wonder Full","Meta AI"],"people":[]},"firstPublishedAt":"2026-11-07T01:50:00Z","updatedAt":"2026-11-07T02:16:05Z","sourceCount":2,"hasPrimarySource":false,"sources":[{"outlet":"Unite.AI","title":"Current Muse Scores 47% on Meta’s Personal Superintelligence Index","url":"https://unite.ai/current-muse-scores-47-on-metas-personal-superintelligence-index","publishedAt":"2026-11-07T02:16:05Z","type":"press","primary":false,"lead":true},{"outlet":"thenewstack.io","title":"Meta introduces specialized versions of its existing models, trained using Muse Wonder, bringing Active Memory and Active Focus to all Muse and Meta AI users","url":"https://thenewstack.io/meta-muse-specialized-models","publishedAt":"2026-11-07T01:50:00Z","type":"press","primary":false,"lead":false}],"sourceNotes":null,"discussions":[],"thread":{"title":"Meta Muse Evolution and AI Agent Race","url":"https://digestai.news/thread/meta-launches-muse-personal-ai-agent-in-singapore","storyCount":4},"cite":{"text":"Digest AI, \"Meta improves Muse and Meta AI with Active Memory and Active Focus\", 7 November 2026, https://digestai.news/story/meta-adds-specialized-muse-models-with-active-memory-and-active-focus","publisher":"Digest AI","title":"Meta improves Muse and Meta AI with Active Memory and Active Focus","datePublished":"2026-11-07T01:50:00Z","url":"https://digestai.news/story/meta-adds-specialized-muse-models-with-active-memory-and-active-focus"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}