Meta Platforms Inc. has entered the AI coding assistant arena with the release of Muse Code, a new AI agent capable of executing complete software engineering tasks across large code repositories. The company announced the beta launch of Muse Code as part of its broader push to regain ground against established AI leaders Anthropic and OpenAI.
Muse Code can plan development changes, write code, and validate outcomes autonomously. It leverages an updated version of Meta’s Muse Spark model, designed specifically to coordinate multiple AI agents working in parallel on complex software projects without interfering with users’ current work. Meta claims that during testing, Muse Code managed simultaneous feature development without conflicts, showcasing its ability to handle extensive, concurrent workflows.
The development of Muse Code reflects Meta’s strategy to capitalize on the increasing adoption of generative AI in software development—a sector that has become a lucrative revenue stream for competitors like Anthropic, with its Claude AI, and OpenAI, known for Codex. The new tool allows users, including those without technical coding skills, to generate functional applications by describing desired features in natural language, streamlining software creation.
Meta’s AI chief Alexandr Wang, who spearheads the company’s Superintelligence Labs, oversees Muse Code’s development as part of a broader initiative to enhance Meta’s AI capabilities following concerns over lagging behind competitors. Wang emphasized affordability as a key differentiator, with pay-as-you-go pricing models and a “contributor tier” that promises to be significantly cheaper than rival offerings, aiming to attract a broader developer base.
This launch comes amid Meta’s recent financial challenges, including a drop in stock value after a disappointing earnings report and lowered revenue expectations. The company’s investment in AI infrastructure and data centers has been substantial, and Muse Code represents a concrete effort to transform those investments into competitive AI products.
By enabling AI agents to work in isolated submodules yet in parallel, Meta hopes Muse Code will improve productivity and reduce coding overhead. The model behind Muse Code, Muse Spark 1.2, orchestrates multiple AI workers to handle different parts of a project simultaneously, an approach designed to avoid code conflicts and increase efficiency.