Meta has officially unveiled its most significant new AI model in years. Dubbed Muse Spark — and previously codenamed Avocado internally — the new model is the first from Meta’s new Superintelligence Labs division and marks a decisive attempt to close the gap with OpenAI, Google, and Anthropic.
Here is everything you need to know about Muse Spark and what it means for the AI landscape.
What Is Muse Spark?
Muse Spark is Meta’s latest AI model, developed by Meta Superintelligence Labs, the AI division that reports to Alexandr Wang, who joined Meta last June as part of a $14.3 billion investment in Scale AI. Wang, who was previously CEO of Scale AI, now oversees Meta’s push to build competitive frontier AI models.
The model is the first entry in Meta’s new “Muse” series. It is now live and powering the standalone Meta AI app and desktop website. Over the coming weeks, Muse Spark will roll out inside Facebook, Instagram, WhatsApp, and Messenger — putting it in front of Meta’s global user base of several billion people. It will also power the AI assistant in Meta’s Ray-Ban smart glasses.
What Can Muse Spark Do?
Meta has highlighted several notable capabilities that position Muse Spark as a direct competitor to frontier models from OpenAI and Google.
Contemplating Mode is one of the most significant new features. When handling complex queries, Muse Spark deploys a squad of AI agents that reason in parallel — a technique Meta says allows it to compete with high-end reasoning modes like Gemini Deep Think and GPT Pro. This parallel reasoning approach is designed to tackle difficult problems that a single AI pass would struggle with.
Shopping Mode is another addition with clear commercial potential. The feature draws from styling inspiration and brand content already circulating across Meta’s apps. It is designed to help users buy clothes, decorate rooms, and discover products through their existing Instagram and Facebook communities.
Meta is also exploring a paid API for Muse Spark, currently in private preview with select partners. The company plans to offer this more broadly at a later date, representing a new revenue stream beyond its traditional advertising business.
Why Meta Needed This
Meta has faced real pressure in the AI market. OpenAI and Anthropic are collectively valued at over $1 trillion, and Google’s Gemini has gained significant traction with consumers and enterprise users alike. Meta’s previous open-source AI models — while widely used — failed to make a meaningful dent in the consumer AI product market.
The company is also spending aggressively. Meta’s AI-related capital expenditure in 2026 is projected at between $115 billion and $135 billion — nearly double what it spent last year. Muse Spark is the product that all of that investment needs to start justifying.
If you have been keeping track of how the major AI assistants compare, our breakdown of Google Gemini vs ChatGPT vs Copilot gives you a solid foundation for understanding where Muse Spark is entering this competitive landscape.
How Does Muse Spark Compare to Gemini and ChatGPT?
Meta has not released formal benchmark comparisons in this announcement, so direct side-by-side performance numbers are not yet available. However, based on the features described, Muse Spark is positioned to compete with:
- GPT-4o and GPT Pro from OpenAI, especially on reasoning and general-purpose conversation.
- Gemini Ultra and Gemini Deep Think from Google, where Meta’s Contemplating Mode is directly aimed.
- Claude from Anthropic, particularly for users who want an AI assistant integrated into consumer social apps.
The Shopping Mode and social integration angle is where Meta has a genuine structural advantage. No other AI lab has direct access to the engagement data and content ecosystems of Facebook, Instagram, and WhatsApp. That distribution advantage could make Muse Spark the most-used AI model in the world very quickly, even if it is not necessarily the most powerful.
For the latest on AI coming to Android specifically, check out our piece on Google Gemini Nano 4 coming to Android.
Frequently Asked Questions
What is Meta Muse Spark?
Meta Muse Spark is a new AI model developed by Meta Superintelligence Labs under Alexandr Wang. It is designed to compete with frontier AI models from OpenAI, Google, and Anthropic, and will power the Meta AI app, Facebook, Instagram, WhatsApp, Messenger, and Ray-Ban smart glasses.
How is Muse Spark different from older Meta AI?
Muse Spark introduces Contemplating Mode, which uses parallel AI agents for complex reasoning, and Shopping Mode, which helps users buy products and discover styles. Previous Meta AI products were powered by third-party models and in-house open-source models. Muse Spark is a new, proprietary frontier model from Meta’s dedicated superintelligence team.
Is Meta Muse Spark available now?
Yes. Muse Spark is currently live in the standalone Meta AI app and desktop website. It will roll out to Facebook, Instagram, WhatsApp, and Messenger in the coming weeks.
Will Muse Spark have an API?
Yes. Meta is developing a paid API for Muse Spark, currently available to select partners in private preview. Broader API access is planned for a later date.
Meta Is Serious About AI Now
Muse Spark is not a minor product update. It is the result of billions of dollars of investment, the hiring of one of Silicon Valley’s most prominent AI executives, and a concerted effort to build a frontier model that can hold its own against the best in the world.
Whether it succeeds in that mission remains to be seen. The AI market is brutally competitive. But with Muse Spark now live and rolling out across Meta’s entire app ecosystem, the company has made its most credible play yet in a race it has been trailing for years.
Leave a Reply