Meta Launches Muse Code AI Agent for Software Development
Meta has launched Muse Code, a terminal-based AI coding agent for software development, powered by its Muse Spark 1.2 model.
SociapaHQ
Meta has officially entered the AI developer tools market with the launch of Muse Code, a terminal-based coding agent that takes on complex, end-to-end software engineering tasks across large codebases.
The tool announced by Meta CEO Mark Zuckerberg is meant to help developers plan, write and validate code in massive software repositories, making it a rival to AI coding agents like OpenAI’s Codex and Anthropic’s Claude Code.
Muse Code is based on Muse Spark 1.2, a code-oriented version of Meta’s flagship AI model, co-trained with the agent framework to accelerate software engineering.
Also Read: Telcos Blocked 1.48 Billion DND-Violating Calls: TRAI
Developer Tools:
Muse Code handles large scale development projects without impacting primary code repositories. It runs multiple sub-agents in parallel using isolated Git worktrees. It enables developers to work on different tasks in parallel, while reducing the chances of conflicts in the shared code base.
The tool also has interactive workflow commands like /plan, /grill and /goal that help developers organise and review coding tasks. Additionally, Muse Code keeps a local append-only event log that provides action traceability and crash recovery.
In vendor-reported benchmark tests, Meta said Muse Spark 1.2 scored 82.9% on Terminal-Bench 2.1 and 59.3% on DeepSWE 1.1, though it still lags Anthropic's Opus 5 on some coding benchmarks.
Also Read: Bisleri Launches Limited-Edition ‘Spider-Man: Brand New Day’ Bottles
Meta has released Muse Code globally through the Meta Model API. Standard pricing starts at $1.25 per million input tokens and $4.25 per million output tokens.
The launch comes as Meta is looking to expand its enterprise AI offerings amid rising competition in the rapidly growing market for AI-powered software development tools. Muse Code aims to further establish its position in the enterprise-grade applied AI space and provide developers with an agent capable of handling more sophisticated coding workflows.
