The plans side by side
Codex is not a standalone product with its own subscription: it comes with the ChatGPT plans. For a team of two or more, ChatGPT Business is the entry point; Enterprise and Edu have their own terms, and some existing workspaces still have Codex-only seats.
| Plan | Price (November 2026) | Codex | For whom |
|---|---|---|---|
| Free | $0 | GPT-6 Luna at Standard speed in the desktop app, subject to rollout | Trying it |
| Go | $8 per month | Same Luna access as Free, for light tasks | One person |
| Plus | $20 per month | Web, CLI, IDE extension and iOS; GPT-6.1 Sol and GPT-6 Luna; credits available | One developer |
| Pro | From $100 per month ($100, $200 or $500) | Everything in Plus; Ultrafast mode on the $500 plan | Heavy individual user |
| Business | $20 per user per month billed annually, $25 monthly, 2+ users | Codex and ChatGPT, larger VMs for cloud tasks, workspace credits | Teams |
| Enterprise and Edu | Contact sales | Everything in Business plus SCIM, EKM, RBAC, audit logs via the Compliance API, retention and residency controls | Large organizations |
| API key | Standard API pricing | CLI, SDK and IDE extension; no cloud features such as GitHub review or Slack | Pay per token |
Business also gives a dedicated workspace with admin controls, SAML SSO and MFA, and OpenAI does not train on that workspace’s data by default.
A worked example
A team of five developers on Business:
- Billed annually: 5 × $20 × 12 = $1,200 per year.
- Billed monthly: 5 × $25 = $125 per month, $1,500 per year.
- On top: credits for anyone who regularly hits their limit.
The misconception: a seat is unlimited
A Business seat is not a blank check. For Plus and standard Business, OpenAI gives estimates of local messages per five hours, and they vary widely by model: 5 to 45 with GPT-6 Astra, 15 to 160 with GPT-6.1 Sol, 35 to 285 with GPT-6.1 Terra, up to 350 to 3,000 with GPT-6 Luna. These are estimates, not fixed caps; large tasks and long sessions use more. Local tasks, cloud tasks and ChatGPT Work draw from the same usage allowance. To see what is left, use /status or the usage dashboard.
Speed costs too. Fast mode uses 2.5 times as much (2 times on purchased credits), Ultrafast on Astra and GPT-6.1 Sol 8 times as much (6 times on credits). Credits are charged per million tokens, at Standard speed: GPT-6.1 Sol costs 50 credits for input, 2.5 for cached input and 250 for output; GPT-6 Luna 2.5, 0.25 and 12.5. What a credit costs in dollars depends on your plan or agreement.
So a team’s real bill is seats plus credits, and the credits depend on which model and speed your developers pick.
What this means for a team
Large tasks and long sessions cost more, so check the usage dashboard before you set a credit budget for the team. Every session also starts by exploring the repo and reading the conventions. Codex memories help one developer work out less from scratch, but they are off by default and stay on their own machine (see Does Codex remember previous work?). A short, good AGENTS.md saves everyone. If part of the team uses Claude Code, the seat price matches: Claude Team costs $20 per seat per month billed annually or $25 monthly. The trade-off is in Codex vs Claude Code.
Where Bombos fits
Bombos does not change what Codex costs; it adds shared context, so a session has less to work out on its own. It is a plugin that gives Codex and the team’s other AI tools one shared memory, with team knowledge, learned rules and what already sits in repos and documents, in the cloud or on your own server. One developer can try it first for their own agents.
Related questions
Is Codex included in ChatGPT Plus?
Yes. Plus costs $20 per month and includes Codex on web, CLI, IDE extension and iOS, with cloud integrations such as automatic code review and Slack. It is a personal plan, without the workspace and admin controls of Business.
What happens when a team member hits their limit?
The current turn can finish, subject to fair use. After that the workspace can buy credits, or the developer can keep working locally with an API key at standard API rates.
Is Codex on an API key cheaper than Business?
It depends on usage: you pay per token with no seat price. You lose the cloud features such as GitHub review and Slack, and models follow what your API key allows.
Is our code used to train models?
According to the pricing page, OpenAI does not train on Business workspace data by default.
