Opinion
Meta Unveils Muse Wonder AI Models for Personal and Professional Tasks
Meta has introduced two new AI models, Muse Wonder and Muse Wonder Full, designed to tackle complex personal and professional tasks. These models are part of Meta's ongoing effort to enhance its AI capabilities and are available through the Meta Model API and Muse Code.
Muse Wonder is set to power Meta's personal assistant, Muse, with subscribers gaining early access. The models boast significant improvements in memory and task execution, featuring a system called Active Memory that retains relevant context across conversations. This allows the AI to remember user preferences and apply them to new tasks effectively.
Another key feature, Active Focus, enables the models to maintain progress toward challenging goals, adapt to new information, and seek clarification from users when necessary. The capabilities of Muse Wonder support various applications, including software development, legal documentation, financial inquiries, and health-related questions.
In terms of performance, Muse Wonder Full achieved a score of 56 on the Artificial Analysis Intelligence Index v4.3, making it the second highest-rated model behind Anthropic’s Claude Opus 5.5. This marks an 8-point improvement over its predecessor, Muse Spark 1.3, which scored 48. Muse Wonder Full also excelled in agentic tasks and coding benchmarks, outperforming competitors like OpenAI’s GPT-6 Astra in several evaluations.
Meta's release of Muse Wonder is part of a broader strategy to reclaim its position in the AI sector, following a significant overhaul of its AI development approach in 2025. The company aims to position its Muse Wonder models as strong contenders against leading AI developers, emphasizing reliability and performance in agentic tasks.
New Developments in AI from Meta
- Meta's CEO Mark Zuckerberg announced the launch of the Muse Wonder AI model, claiming it is now competitive with top models from OpenAI and Anthropic.
- The Muse Wonder model is developed by Meta Superintelligence Labs and is designed to excel in tasks such as coding and professional work.
- This announcement follows previous challenges faced with Meta's Llama 4 model, marking a significant turnaround for the company in the AI space.
- Market data shows a decrease in confidence regarding Anthropic's model, with predictions suggesting it may not be the top performer by October 2026.
- Google's AI model currently leads the market with a 70.9% probability of being the best-performing model.
- Future benchmark updates and leaderboard changes may further affect AI model rankings, with potential shifts expected if Muse Wonder achieves independent verification of its capabilities.
New Developments on Meta's Muse AI
- Launch Date: Meta launched Muse on September 8, 2026.
- Functionality: Muse is designed to act as a personal AI agent that can perform tasks such as booking travel, sending emails, and making purchases through Stripe Link.
- User Age Requirement: Muse is available only to users aged 18 and older, with age verification required.
- Subscription Tiers: The base version of Muse is free with usage limits, while paid tiers include Power at $20/month and Maximum at $100/month.
- Privacy Commitment: Meta has emphasized a commitment to user privacy, stating that user data will not be shared with advertisers.
- Interactive Features: Muse includes an interactive avatar with customizable features and provides each user with a dedicated cloud virtual machine for their agent's tasks.
- Policy on Explicit Content: Muse's official policy bans adult content, erotic chat, and pornography, with no exceptions made for age verification.
FAQ
What are Muse Wonder and Muse Wonder Full?
Muse Wonder and Muse Wonder Full are two new AI models introduced by Meta, designed to handle complex personal and professional tasks, enhancing the capabilities of Meta's AI offerings.
How do the Muse Wonder models improve task execution?
The Muse Wonder models feature a system called Active Memory, which retains relevant context across conversations, allowing the AI to remember user preferences and apply them effectively to new tasks.
What applications can the Muse Wonder models support?
Muse Wonder models can support a variety of applications, including software development, legal documentation, financial inquiries, and health-related questions.
How do Muse Wonder models perform compared to other AI models?
Muse Wonder Full achieved a score of 56 on the Artificial Analysis Intelligence Index v4.3, making it the second highest-rated model, and it outperformed competitors like OpenAI’s GPT-6 Astra in several evaluations.
What is Meta's goal with the release of Muse Wonder models?
Meta aims to reclaim its position in the AI sector with the Muse Wonder models, emphasizing reliability and performance in agentic tasks to compete against leading AI developers.