OpenAI introduces Codex-only seats with $0 fixed cost and usage-based team pricing

• OpenAI now lets ChatGPT Business and Enterprise customers add Codex-only seats with no fixed seat fee, with billing based on token usage instead of a bundled license OpenAI launch post Official pricing link.
• Those Codex-only seats have no rate limits, while standard ChatGPT Business seats still include Codex usage limits, according to OpenAI’s pricing announcement Launch tweet Official pricing link.
• OpenAI also cut annual ChatGPT Business pricing from $25 to $20 per seat and is offering eligible workspaces $100 in credits per new Codex-only user, up to $500 total Credit offer Official pricing link.
• The company says Codex usage inside Business and Enterprise is up 6x since January, and more than 2 million builders now use Codex weekly Official pricing link Commentary summary.
• OpenAI is pairing the pricing change with a push around the desktop app, Plugins, and Automations, and one OpenAI employee said the Codex app is now the product’s most-used surface, ahead of the VS Code extension and CLI gdb on trying Codex at work Most-used surface claim.

OpenAI’s pricing post is unusually direct: zero fixed-cost Codex seats, no rate limits, and a lower Business seat price in the same update. The Codex product page fills in the product push around that pricing, with built-in worktrees, cloud environments, Skills, and Automations. There is also a separate Codex offer page carrying the same up-to-$500 credit pitch now showing up across the launch tweets.

OpenAI’s main change is simple: teams can now add users whose seat is only for Codex, without paying a fixed per-seat fee first.

The official post says those seats are available to ChatGPT Business and Enterprise workspaces, with usage billed on token consumption rather than a flat license. That shifts Codex from a bundled perk inside ChatGPT plans into something teams can pilot with narrower budget ownership, as summarized in the launch post and echoed by early commentary Commentary summary.

The fine print that matters is how OpenAI split access paths.

• Codex-only seats: no rate limits, usage billed on token consumption.
• Standard ChatGPT Business seats: broader ChatGPT access, but Codex usage limits still apply.
• Annual ChatGPT Business pricing: reduced from $25 to $20 per seat in the same announcement.

That combination makes the product menu a lot more explicit than before. The official post frames it as easier cost tracking across budgets, workflows, and teams, not just a cheaper way to buy seats Pricing post.

OpenAI attached a temporary subsidy to the rollout: eligible ChatGPT Business workspaces can get $100 in credits for each new Codex-only team member, capped at $500 per team.

The same post pairs that incentive with a bigger usage claim. OpenAI says more than 9 million paying business users rely on ChatGPT for work, more than 2 million builders use Codex every week, and the number of Codex users inside Business and Enterprise has grown 6x since January. It also names Notion, Ramp, Braintrust, and Wasmer as existing team deployments Pricing post.

OpenAI is not selling metered seats into a blank product. The pricing post points teams first to the macOS and Windows Codex app, then calls out Plugins and Automations as the connective tissue for existing systems.

The product page adds more concrete detail:

• The Codex app is described as a command center for multi-agent coding.
• It includes built-in worktrees and cloud environments for parallel agent work.
• Skills are positioned as reusable team standards for code understanding, prototyping, and documentation.
• Automations cover background tasks like issue triage, alert monitoring, and CI/CD.
• The same account spans the app, editor, and terminal surfaces.

Those details make the pricing move look less like a pure billing tweak and more like a push to get entire teams onto Codex as a standalone engineering surface Codex product page.

One of the more revealing launch-day details came from OpenAI engineer Thomas Sottiaux, who said the Codex app is now the company’s most-used surface, ahead of both the VS Code extension and the CLI.

That gives the desktop app a different weight in this rollout. The official pricing post recommends starting there, and the public product page already treats the app as the first surface before editor and terminal handoff Codex product page. If that usage ranking holds, OpenAI is no longer talking about Codex primarily as an extension or terminal tool. It is talking about a desktop control plane for multi-agent coding.