Meta Muse Scores 47% on Personal Superintelligence Index

By Daniel Okafor

Meta AI Research said on November 6, 2026 that the current Muse model scores 47% on Meta’s Personal Superintelligence Index, compared with 38% for the previous model.

The company is rolling out specialized versions of the models behind Muse and Meta AI. Meta said those versions were trained using Muse Wonder and improve the existing systems rather than constitute a rollout of the Wonder series itself. The approach extends capabilities introduced with Wonder to more people through the assistants’ existing models.

What changes in the November rollout?

Meta’s October 12, 2026 announcement established the sequence. The company described Muse Wonder and Muse Wonder Full as its “most capable models to date.” Wonder would first power Muse experiences for subscribers before being used to develop specialized models for everyone using Muse. The November rollout carries that specialized-model plan into both Muse and Meta AI.

How does Active Memory decide what to keep?

Meta said Active Memory lets the assistants “retain important information and retrieve it when it becomes relevant to a later conversation.” Its training covers the difference between temporary and permanent information. It also covers reasoning over time and recognizing when details remain relevant or have become stale.

Meta described memory as a selection and timing problem. The model must identify which earlier context applies now and which has become stale.

For the improved Muse models, Meta reports a 2.4x reduction in forgetting information critical to users’ goals. Its Active Memory Eval shows the current model correctly using critical context in 90% of tasks, compared with 76% for the previous model.

Meta’s results describe two separate parts of the problem: retaining information needed for a goal and applying that context correctly during a task. Together, the figures point to gains in both retaining context and deploying it within a task.

What does Active Focus change when work hits a snag?

Active Focus is intended to keep Muse moving toward a goal when an unexpected challenge interrupts the work. Meta said it trained the models on varied bottlenecks designed to reflect real-world conditions. The training teaches the systems to persist, recover, or pursue alternatives for the user.

In Muse, Meta reports a 67% improvement in completing everyday tasks where unexpected issues arise.

Meta links persistence to responsiveness. The specialized models are designed to “multitask, choose between independent progress and live collaboration,” Meta said. They can also incorporate steering or clarifications without becoming confused or losing the task. That lets an agent continue pursuing the goal while changing course when the user intervenes.

What changes for Meta AI?

In Meta AI, Meta said the same improvements are intended to make the assistant a better thinking partner and collaborator. The assistant can divide complex problems into clear steps, ask sharper follow-up questions, and keep pace when a user changes direction mid-conversation.

In practical terms, Meta’s design gives memory and focus separate jobs. Memory decides which earlier details still apply. Focus helps the system recover when a task changes or stalls. The practical test resembles a long working session: useful details must survive interruptions, and a correction should redirect the work rather than reset it.

What will determine whether the rollout works?

For users, Meta frames the test as continuity. Three outcomes matter:

• Useful context survives changing circumstances.
• Work recovers from obstacles.
• Corrections enter the process without losing the original goal.

Meta’s reported gains connect those outcomes to the everyday collaboration it wants to support in Muse and Meta AI. The next test is whether the specialized models can preserve that behavior outside the reported evaluations and in users’ changing work.