Meta rises after Muse Spark 1.3; BofA sees 32% upside

Meta Platforms shares rose about 4% after the company released Muse Spark 1.3, a model designed for coding and longer-running agentic tasks. The stock move followed publication of performance details from the company.

Meta reported that Muse Spark 1.3 uses roughly 20% fewer tool calls and about 25% fewer tokens than Muse Spark 1.2 on comparable engineering tasks. The company said the model handles longer-horizon work more effectively, manages multiple workflows within a single thread and improves coding efficiency. Muse Spark 1.3 is available through Muse Code and the Meta Model API.

Bank of America analyst Justin Post reiterated a Buy rating and an $810 price target, implying about 32% upside from Meta’s Sept. 3 close of $610.68. Post cited Meta’s rapid model-release cadence and the agentic improvements as relevant to the company’s internal consumer AI project, Hatch. BofA estimates that deployments of Meta’s planned MTIA custom chips could represent 15% to 20% of the company’s AI capacity and could lower compute costs as workloads grow. BofA based its $810 target on 24 times projected 2027 GAAP earnings; at the time of the note Meta traded near 18 times projected 2027 GAAP earnings, below its historical multiple of roughly 21 times.

Bernstein reiterated an Outperform rating with an $800 price target, writing that model improvements can boost recommendations, ad targeting and engagement across Meta’s apps and feed into advertising revenue. KeyBanc cut its price target to $760 from $855 while keeping an Overweight rating, describing progress at Meta Superintelligence Labs on Muse Spark as “meaningful” and saying investors still need clearer proof that technical advances convert into wider customer adoption and revenue gains.

Analysts and investors highlight valuation and cost dynamics as central issues. The company has increased AI infrastructure spending, which raises fixed costs and can pressure margins and free cash flow unless model and chip improvements produce measurable cost savings or additional revenue.

Market participants will watch for signs of adoption: greater use of Muse Spark by developers and advertisers, practical consumer-facing agents such as Hatch, and evidence that MTIA chips reduce computing costs. Muse Spark 1.3’s efficiency gains provide new, specific metrics for those assessments.