Meta releases coding agent to compete with OpenAI, Anthropic

By the Editorial Team · Friday, August 7, 2026

Meta Platforms launched Muse Code, a new AI-powered coding agent aimed at providing a more affordable alternative to competitors like OpenAI and Anthropic. Priced significantly lower, Muse Code is expected to gain wider internal use at Meta and help the company capitalize on its AI investments amid investor pressure for returns.

Meta Platforms unveiled a new coding agent, Muse Code, on Wednesday, entering the competitive market for AI-powered software development tools. The company positions Muse Code as a more cost-effective alternative to existing products from rivals such as OpenAI and Anthropic.

Alexandr Wang, Meta’s head of AI and co-founder of ScaleAI, emphasized the potential for Muse Code to offer significant value across various workflows, particularly due to its pricing structure. He noted that some Meta employees are already utilizing the agent internally and anticipated broader adoption within the company over time.

Coding agents are artificial intelligence tools designed to assist users in software engineering tasks without requiring them to write code manually. They have become a focal point for Silicon Valley firms aiming to streamline programming and accelerate development. Anthropic’s Claude Code gained traction last December, helping the company establish a strong position in the AI sector, while OpenAI markets its Codex coding agent under subscription plans.

Muse Code features two pricing tiers: one aligned with Meta’s general Muse Spark model, and a substantially lower-cost option priced at 20 cents per million output tokens—a unit measuring AI computational work. To access this cheaper tier, users must consent to provide feedback to improve the system. By comparison, pricing for agents from companies such as Anthropic and OpenAI typically runs around $20 monthly, with pay-as-you-go charges ranging from $10 to $30 per million tokens depending on the model. Meta’s prices are also competitive with those of Chinese AI coding tools like DeepSeek, which can cost as little as $18 per million output tokens.

Meta’s introduction of Muse Code comes amid pressure from investors to demonstrate tangible returns on the company’s considerable investment in artificial intelligence infrastructure and talent acquisition. Shares of Meta dropped 10 percent last week following a quarterly earnings report in which Chief Executive Mark Zuckerberg declined to provide new information about the company’s plans for a cloud computing service.

Last year, Meta restructured its AI initiatives by establishing Meta Superintelligence Labs, with Wang leading the effort. The division has aggressively recruited experts with multimillion-dollar compensation packages to bolster the company’s AI capabilities and sharpen its competitive edge in the evolving technology landscape.

This article was reported by synthesizing coverage from multiple independent news outlets to present a balanced, unbiased account. Read about our editorial standards. Spotted an error? Let us know.