By Meghan Bobrowsky and Tina Li

Meta Platforms wants in on Silicon Valley's hottest product du jour: coding agents.

The company on Wednesday announced a new coding agent, called Muse Code, that it is pitching as a cheaper alternative to competitors.

"We think that for a lot of workflows and a lot of use cases, this can be an incredibly good option, especially from a cost perspective," Alexandr Wang, Meta's AI chief, said in an interview. He said some Meta employees are already using it internally and that he expects internal usage to ramp up.

Investors have pressed Meta to show returns from its lavish spending on AI talent and computing power. Shares in the company fell 10% last week following a quarterly earnings call in which Chief Executive Mark Zuckerberg declined to provide new details about the company's discussions around establishing a cloud-computing service.

Coding agents allow AI users to develop projects and complete software engineering work without having to write code themselves. Anthropic's Claude Code took off in December, powering the company to a leading position in the AI race, and OpenAI released its own offering, called Codex.

Meta's Muse Code agent has two price tiers: one that's the same as its general Muse Spark model and another that is more than 10 times cheaper. To access the less-expensive tier, which costs $0.20 per million output tokens, users must agree to provide feedback to help improve the agent. Tokens are the basic unit of artificial-intelligence computing.

Claude Code and Codex come bundled in pricing plans that cost roughly $20 a month, with more expensive plans for increased usage. Exceeding the usage cap switches the user to pay-as-you-go rates. Rates per million output tokens range from $12 for GPT-5.6 Terra and $30 for Sol, with Claude Sonnet 5 at $10 and Opus 5 at $25.

Meta's new coding agent is priced roughly on par with models from China, such as DeepSeek, that cost as little as $0.18 per million output tokens. OpenAI has also slashed prices on older models, such as GPT-5.6 Luna, which dropped from $6 to $1.20 per million tokens.

Meta overhauled its AI efforts last summer and created a new division called Meta Superintelligence Labs. Wang, the co-founder of ScaleAI, was put in charge of it.

He embarked on an aggressive hiring spree, using multimillion-dollar pay packages to attract talent from leading labs, and began development of Muse Spark, the new AI model Meta released in April.

Write to Meghan Bobrowsky at meghan.bobrowsky@wsj.com and Tina Li at tina.li@wsj.com

(END) Dow Jones Newswires

August 05, 2026 15:45 ET (19:45 GMT)