Following the announcement, Meta’s stock surged as much as 9.5% intraday and ultimately closed up 6.5% at $612.42, with trading volume jumping over 250% from the previous session. On the same day, news of a U.S.–Iran ceasefire fueled a broader tech rally, with all three major U.S. indices rising more than 2%.

Understanding Muse Spark

Muse Spark represents a reset for Meta in the AI race rather than a simple upgrade to its Llama series. According to MSL head Alexandr Wang, the team has “rebuilt the entire AI stack from scratch” over the past nine months — including new infrastructure, architecture, and data pipelines — with Muse Spark as the result.

The model natively supports multimodal understanding and integrates tool usage, visual reasoning chains, and multi-agent coordination. It also offers three reasoning modes: “instant,” “think,” and “deep think.” The deep reasoning mode can coordinate multiple sub-agents for parallel reasoning, directly competing with Google’s Gemini Deep Think and OpenAI’s GPT-5.4 Pro.

In terms of efficiency, Meta claims that Muse Spark achieves comparable reasoning performance using only one-tenth of the compute required by Llama 4 Maverick. More importantly, Muse Spark adopts a closed-source approach — neither model weights nor architecture are publicly released — with future monetization likely through APIs or subscription models. This marks a significant shift, as Meta moves from an open-source AI advocate to a competitor focused on proprietary technology and commercial returns.

Meta has taken a pragmatic approach in positioning the model. In the Artificial Analysis Intelligence Index v4.0, Muse Spark ranks fourth with a score of 52, behind Gemini 3.1 Pro (57), GPT-5.4 (57), and Claude Opus 4.6 (53). The gap is particularly noticeable in advanced reasoning benchmarks such as ARC AGI 2.

However, Muse Spark performs strongly in healthcare-related reasoning tasks. Meta collaborated with over 1,000 doctors to curate training data, achieving scores of 78.4 on MedXpertQA and 42.8 on HealthBench Hard — outperforming some competitors.

Another standout feature is its shopping mode. The model can generate personalized recommendations based on user preferences and interactions across Instagram, Facebook, and Threads, potentially enhancing e-commerce conversion and advertising effectiveness.

Three Layers of “Certainty” Driving Meta’s Stock

The first is timing certainty. Markets had been waiting for tangible output from Meta’s Superintelligence Lab. Reports indicated that the earlier Avocado model underperformed in internal testing, leading to delays and lowered expectations. The launch of Muse Spark finally delivered a long-awaited milestone.

The second is business model clarity. Bank of America analyst Justin Post described the launch as removing a key overhang on the stock, reiterating a Buy rating with an $885 price target. Investors see Muse Spark as a critical driver for boosting engagement and monetization across Meta’s platforms.

The third is improving competitive dynamics. At the same time as Muse Spark’s release, OpenAI shut down its short-form video generation tool Sora, reducing competitive pressure on Meta’s short video advertising business. Combined with a broader tech rally driven by geopolitical easing, this provided a favorable macro backdrop.

Meta shares remain down 7.2% year-to-date and more than 22% below their 52-week high of $796.25 reached last August. For investors, the valuation framework is shifting toward viewing Meta as an AI-driven content and advertising platform.

Going forward, the key questions will be whether Muse Spark can deliver measurable improvements in e-commerce conversion, ad targeting accuracy, and user retention over the coming quarters — and whether future iterations can narrow the technological gap with Google and OpenAI.