Accra:Meta has introduced two new artificial intelligence models, Muse Wonder and Muse Wonder Full, in a strategic move to compete more effectively with leading AI systems developed by OpenAI, Anthropic, and Google.
According to Anadolu Agency, the newly launched models are Meta's most advanced AI offerings to date. They are designed for use in both personal and professional AI applications, capable of executing long-running tasks, utilizing computers, and functioning in specialized fields such as law, finance, and health. These models are also equipped to manage complex coding projects, delegate tasks to smaller AI agents, and recover from errors without losing sight of their original objectives.
One of the standout features of these models is Active Memory, which enables the AI to retain and recall important user information across multiple interactions. Additionally, the Active Focus feature is designed to keep AI agents operational even when challenges arise, by encouraging them to try alternative methods, adapt to new data, and seek clarification when necessary.
Meta has made Muse Wonder accessible to developers via the Meta Model API and Muse Code programming tool, while the more advanced Muse Wonder Full is available through a subscription service starting at $50 per month. Initially, Muse Wonder will be integrated into Meta's Muse personal assistant for subscribers.
The company has released benchmark results indicating that these models perform on par with or better than some systems from OpenAI and Anthropic in selected areas such as coding, memory, and professional tasks. However, further independent testing is necessary to fully evaluate the models' overall performance.
Meta has emphasized the importance of privacy and security, stating that the Wonder family of models is trained to safeguard private information, reject harmful activities, and bolster defenses against cyber threats. A comprehensive safety report is expected to be published in the future.
This launch is part of Meta's broader initiative toward developing AI systems that align with CEO Mark Zuckerberg's vision of 'personal superintelligence,' which aims to understand users' goals, preferences, and long-term needs. Unlike some previous models, Meta has not announced plans to release the Wonder models as open source.