Meta has unveiled two new artificial intelligence models as it seeks to compete with the world’s leading AI companies. The models, called Muse Wonder and Muse Wonder Full, are designed to support personal and work-related AI agents.
The company announced the models on Monday and said they were available immediately. Meta claims they can deliver top-level results across a range of AI tasks. The launch marks another step in its plan to become a stronger force in advanced AI.
The new models are part of Meta Superintelligence Labs, a major AI project backed by large investments. They are expected to power more features in the company’s Muse AI agent, which launched early last month and has quickly gained users.
Meta says the models include two key features called Active Memory and Active Focus. Both are designed to make AI assistants more useful and better able to handle tasks over time.
Active Memory allows an AI assistant to recall useful details from earlier conversations. This means users may not need to repeat the same information each time they start a new chat. The feature could make AI tools more personal and save time for people who use them regularly.
Active Focus helps an AI system keep working when it faces errors or other problems. Meta says the models can continue trying to complete a task instead of stopping at the first roadblock. They can also ask users for more details or permission before taking certain actions.
These features reflect a wider push across the technology industry to build AI agents that can do more than answer questions. Companies want their tools to remember user needs, plan steps and carry out tasks with less direct guidance.
Meta says Muse Wonder and Muse Wonder Full can compete with leading systems from other major AI developers in some benchmark tests. The company named Anthropic’s Fable 5.1, OpenAI’s GPT-6 Astra and Google’s Gemini 4 Argon among the models it believes its new products can challenge.
However, the claims relate to selected benchmarks. They do not prove that Meta’s models outperform rivals across every task or real-world use. Independent testing and wider user experience will help show how well the new systems perform.
The launch is a notable change for Meta, which faced questions about its position in the AI race last year. Its latest products suggest the company is working to close the gap with competitors that have gained attention for advanced chatbots and AI services.
That progress has required substantial spending. Meta has invested tens of billions of dollars in its AI plans. One major deal was its $14 billion investment in Scale AI, which helped bring the company’s co-founder and chief executive, Alexandr Wang, into Meta as its chief AI officer.
Meta has also recruited AI researchers and other experts from rival technology companies. The hiring drive is part of its effort to build the teams and systems needed to develop more advanced models.
Yet the spending has raised concerns among investors. Building AI systems requires major costs for computing power, data centres, research and skilled staff. Investors want to know whether these investments will lead to stronger revenue and profits over time.
Meta’s share performance has also lagged behind some major technology rivals. Its stock rose about 6% over the previous 12 months, according to the figures cited in the report. Microsoft gained about 5% during the same period, while Amazon rose roughly 20% and Google climbed around 45%.
Those results have added pressure on Meta to show that its AI strategy can deliver lasting business value. New products may help attract users and create fresh ways to earn money, but the company still needs to prove their financial impact.
The launch of the Muse AI agent and plans for the Muse Charm device have helped improve the outlook for Meta’s AI efforts. The device is expected to reach the market later this year. Together, these projects show the company’s aim to bring AI into more parts of daily life.
The next test will be whether users continue to adopt Meta’s AI tools and whether those products can support future growth. Strong benchmark results may help build confidence, but investors will also watch user numbers, product demand and costs.
Meta’s latest models give the company a stronger platform to compete in advanced AI. Still, the long-term success of its strategy will depend on turning technical progress and heavy investment into clear financial returns.
