Meta's Muse Spark handles text, images and audio across WhatsApp, Instagram and Facebook, scoring 52 on the Intelligence Index after a March delay.
Meta shipped Muse Spark last week, ending a delay that had stretched from March and sending its stock nearly 10% higher across five trading sessions. The model, originally codenamed Avocado inside the company, is now live on WhatsApp, Instagram, Facebook, Messenger, and Meta's AI glasses hardware. Shares closed Friday at $629.86.
Muse Spark is the debut product from Meta Superintelligence Labs, a unit the company assembled as its frontier AI push. It scores 52 on the Intelligence Index and handles multimodal reasoning, including health-related queries, on surfaces that collectively reach billions of daily users.
The delay that wasn't a disaster
Meta pulled Avocado from its March rollout after internal tests showed it underperforming on key benchmarks. According to AOL Finance, the additional weeks allowed the team to close the gap with rivals rather than release a model that would land below expectations. The call now looks defensible: a late model that competes is a better outcome than a punctual one that gets demolished in public benchmarks.
The competitive field is genuinely crowded. 9to5Google reported this week that Gemini 3.1 Pro is already reaching subscribers across multiple subscription tiers, and OpenAI's GPT-5.4 has been shipping for weeks. Both rivals have head starts on user adoption. The broader shift is worth noting: frontier labs that once raced to ship now treat internal quality thresholds as hard launch gates, absorbing schedule slippage rather than releasing models that will embarrass them in third-party evaluations.
Valuation math
Meta trades at 18x trailing earnings of $23.98 per share, below Alphabet's 23x and Microsoft's 19.6x. Revenue for 2025 came in at $201 billion, up 22% year over year, giving the company more room than most to sustain data center spending and compete for AI talent. That valuation gap to peers implies the market has not fully priced in AI monetization upside, which is exactly what Muse Spark is supposed to begin delivering.
Those numbers frame what the stock move actually signals. A 10% gain in five sessions on a model release is unusual; it suggests investors believe AI will translate into measurable revenue, not just usage metrics. Whether Muse Spark produces that conversion, through ad revenue lift or new subscription flow, is the central open question heading into earnings.
Distribution vs. architecture
The health task capability deserves specific attention. Consumer AI handling medical reasoning faces real exposure: liability questions around clinical guidance, potential FDA scrutiny, and the trust deficit that tends to follow high-profile errors in a health context. Meta has not released a detailed policy on how Muse Spark handles health queries or whether it routes users toward verified sources.
Meta's distribution advantage is real regardless of those concerns. Muse Spark arrives inside apps that people open dozens of times daily, a structural edge no API-first competitor can replicate quickly. AOL Finance attributed part of the stock surge to this logic, noting that reach matters as much as benchmark score. History supports the argument: the company controlling the interface has routinely outperformed the company with the better underlying model, though Meta will need to prove that pattern holds when users have access to 9to5Google-covered alternatives from Google and OpenAI a single tap away.
Analysts will push Meta on revenue specifics at the next earnings call. A strong launch-day stock move is a start. Sustaining it requires Muse Spark to appear in concrete monetization metrics, not just engagement benchmarks. If the company cannot point to measurable impact by mid-year, the market will revisit whether those extra months of development justified the wait.
FAQ
What is Muse Spark?
Meta's first major AI model in more than a year, developed by Meta Superintelligence Labs. It handles multimodal reasoning including health queries and runs natively on WhatsApp, Instagram, Facebook, Messenger, and Meta AI glasses.
Why did Meta delay Muse Spark from its March launch?
Internal tests showed the model, then codenamed Avocado, underperforming against competitors. Meta chose to hold the release until the team could narrow the gap with rivals including Gemini 3.1 Pro and GPT-5.4.
How does Muse Spark compare to Gemini and GPT-5.4?
Muse Spark scores 52 on the Intelligence Index. Gemini 3.1 Pro and GPT-5.4 are already in the market with established user bases. Meta has not published a direct head-to-head benchmark comparison.
Is Meta stock cheap compared to other AI companies?
At 18x trailing earnings, Meta trades at a discount to Alphabet at 23x and Microsoft at 19.6x, despite posting $201 billion in 2025 revenue growing 22% year over year. The launch appears to have partially closed that gap.