Meta Launches Muse Spark AI: New Features Arrive for Facebook and Instagram
Meta has officially announced Muse Spark, the first in a new series of AI models built by its Superintelligence Labs. Developed over the last nine months, the model marks a total rebuild of the company’s AI stack, designed to be faster and more capable of reasoning through complex questions in science, math, and health.
More importantly, this is the first AI model launched from Meta after Mark Zuckerberg and team spent billions poaching talent from other companies.
The new model currently powers the Meta AI app and website and will roll out to WhatsApp, Instagram, Facebook, Messenger, and AI glasses in the coming weeks. A key change is the introduction of subagents that can work in parallel. For example, when planning a trip, different agents can simultaneously draft an itinerary, compare destinations, and find activities to provide a faster, more complete answer.
Muse Spark features strong multimodal perception, meaning the AI can now see and understand photos rather than just reading text. Users can snap a photo of a snack shelf to identify high-protein options or scan products to compare them to alternatives. This capability will also be integrated into Meta’s AI glasses to help the assistant understand the physical world around the wearer.
The update introduces several specialized capabilities, including a health feature where Meta worked with physicians to help the AI navigate health concerns and understand images or charts. Also, a new visual coding feature allows users to create custom websites and mini-games straight from a text prompt. For those looking for inspiration, a new shopping mode surfaces styling ideas and product recommendations based on the creators and communities already active across Meta’s apps.
It’ll be interesting to see how well Muse Spark performs. Most people seem to have their favourite AI chatbot and it’s hard to switch away if you’re locked in.
While Muse Spark is the first model in the series, Meta confirmed it is a foundation for larger models already in development. As the rollout expands to more countries, the AI will begin weaving Reels, photos, and public posts directly into answers to provide more relevant context from local communities.
Want to see more of our stories on Google?
P.S. Want to keep this site truly independent? Support us by buying us a beer, treating us to a coffee, or shopping through Amazon here. Links in this post are affiliate links, so we earn a tiny commission at no charge to you. Thanks for supporting independent Canadian media!

