1M Downloads Later, OpenAI Hints Codex’s Free Ride Won’t Last

Codex is having a “ChatGPT moment.”

OpenAI’s new standalone desktop app for Mac, which launched on Feb. 2, has officially crossed the 1 million download mark in its first seven days. The rapid surge in users marks a significant milestone in the “AI coding wars,” signaling that developers are moving quickly to adopt tools that not only suggest code but also build it.

Unlike the standard chat interface many are used to, the Codex app is being pitched as a “command center” for what experts call agentic coding. Powered by the GPT-5.3-Codex model, the app enables users to run multiple AI agents simultaneously. While one agent hunts down bugs, another can write a new feature or run tests in the background.

On X, OpenAI CEO Sam Altman revealed the scale of the growth: “More than 1 million people downloaded Codex App in the first week. 60+% growth in overall Codex user last week!”

Free access won’t last forever

The massive download numbers were fueled in part by a limited-time promotion that allowed ChatGPT Free and Go users to try the app. However, as the milestone was reached, Altman hinted that wide-open access might soon face restrictions to manage the high costs of operating such a powerful model.

“We’ll keep Codex available to Free/Go users after this promotion; we may have to reduce limits there but we want everyone to be able to try Codex and start building,” Altman wrote on X.

For now, paid subscribers (Plus, Pro, Team, and Enterprise) still enjoy the highest usage limits, but those on the free tiers should expect some throttling soon.

A hot market gets hotter

Codex success proves there’s a massive appetite for advanced, agentic coding tools. But OpenAI isn’t the only player at the table.

As eWeek previously reported, competitors are seeing serious traction. Anthropic recently announced that its Claude Code product reached $1 billion in annualized revenue remarkably quickly. Meanwhile, other tools take a model-agnostic approach, allowing developers to use a variety of AI models through a single interface rather than being locked into a single ecosystem.

For developers and tech leaders, this means choice is expanding rapidly. The focus is shifting from which tool can write a line of code to which platform can best manage a squad of AI workers safely and efficiently.

Also read: The best AI detection tools in 2026 highlight how quickly verification software is evolving alongside mainstream generative AI.