Meta released Muse Spark on Wednesday, marking the debut of its superintelligence team assembled through a $14.3 billion talent war that included hiring Scale AI CEO Alex Wang and offering engineers pay packages worth hundreds of millions of dollars.
The model represents Meta's attempt to close the gap with AI leaders OpenAI and Google. Meta shares jumped nearly 7% following the announcement as investors evaluated whether the company's massive AI investments would generate returns.
Muse Spark will initially power Meta's AI app and website before replacing existing Llama models across WhatsApp, Instagram, Facebook and Meta's smart glasses in coming weeks. The model belongs to an internal family codenamed Avocado — the first in Meta's new Muse series.
This initial model is small and fast by design, yet capable enough to reason through complex questions in science, math, and health. It is a powerful foundation, and the next generation is already in development
Meta — company blog post
The company designed Muse Spark for everyday tasks rather than technical applications. Users can photograph airport snacks to identify protein-rich options, estimate meal calories from images, or visualize how objects would look in different settings.
German tech publication emphasizes Meta's strategic positioning against competitors and provides technical analysis of the model's capabilities. The outlet frames this as part of the broader US tech competition while noting performance limitations in programming — reflecting Germany's focus on technical precision and industrial competitiveness in AI development.