Meta Debuts First AI Coding Agent To Take On Anthropic and OpenAI

Meta unveiled Muse Code, its inaugural AI coding agent, asserting it serves as a more budget-friendly alternative to Anthropic's Claude and OpenAI's Codex models. CNBC highlights that this solution features pay-as-you-go pricing and an optional zero-data-retention option for enterprise users. Meta's chief artificial intelligence strategist, Alexandr Wang, spearheads Meta Superintelligence Labs and manages foundation model development.

Wang joined the company in June of the previous year as part of CEO Mark Zuckerberg's initiative to revamp Meta's struggling AI strategy. According to Wang, "You can install it with a single command and then leverage it to tackle complete software engineering tasks across a broad spectrum of use cases, including planning changes, composing code, and verifying outcomes."

The new Muse Code tool, accessible in a preview version, functions seamlessly within Meta's latest AI model, Muse Spark 1.2. Wang refrained from disclosing user statistics for Meta's Muse Spark AI models, but stated that "adoption has been exciting and robust." Wang noted that the development and training of the Muse Spark model took place concurrently with Muse Code implementation, enhancing overall coding performance.