{Economic Information Daily, 13th} Meta (US.META) unveiled its latest artificial intelligence model, Muse Wonder, on Monday (12th), focusing on enhancing the execution capabilities of personal and professional AI agents to assist users in handling complex tasks such as bookings, programming, legal, and financial matters, further expanding revenue sources beyond its core online advertising business.
Meta stated that Muse Wonder, codenamed "Watermelon," features "active memory" and "active focus" functions, capable of retaining important information across conversations and attempting alternative methods when tasks are blocked. The model is available to developers starting today via the Meta Model API and Muse Code, and will later be accessible to paying users of the Muse personal AI agent application.
The Muse application offers two subscription tiers at $20 and $100 per month; the higher-tier Muse Wonder Full is available to Muse Code subscribers at $50 or $200 per month. Meta claims the new model has achieved competitive performance in various agent tasks and programming benchmark tests, though its actual capabilities still require further independent validation.
Meta also noted that the new model includes enhanced security protections, but as AI agents involve access to emails, messages, and personal data, privacy and authorization management remain focal points of market concern. (kk)
Meta stated that Muse Wonder, codenamed "Watermelon," features "active memory" and "active focus" functions, capable of retaining important information across conversations and attempting alternative methods when tasks are blocked. The model is available to developers starting today via the Meta Model API and Muse Code, and will later be accessible to paying users of the Muse personal AI agent application.
The Muse application offers two subscription tiers at $20 and $100 per month; the higher-tier Muse Wonder Full is available to Muse Code subscribers at $50 or $200 per month. Meta claims the new model has achieved competitive performance in various agent tasks and programming benchmark tests, though its actual capabilities still require further independent validation.
Meta also noted that the new model includes enhanced security protections, but as AI agents involve access to emails, messages, and personal data, privacy and authorization management remain focal points of market concern. (kk)
