Codex for a Team: Per-Seat Billing Plus Overages Is Why the Bill Spirals
The spiral has a specific anatomy: ChatGPT Business Standard seats ($20/user annual) carry Plus-level Codex limits per person (Premium seats get 5x and no five-hour limit), so daily-driver devs blow through them, overflow goes to credits billed per model (GPT-6 Astra at ~50x GPT-5.6 Luna rates), and anything running on raw API keys is unmetered by anything except the invoice. Fixing the number means separating the two workloads: interactive Codex stays on seats (that part is genuinely fixed), and the volume work (long agentic runs, background tasks) moves to a custom provider with a hard budget. Codex CLI supports that natively via [model_providers] in config.toml, which is how a team of five gets one predictable line item instead of five surprise ones.
Where the money actually goes
| Component | Billing shape | Predictability |
|---|---|---|
| ChatGPT Business seats | $20/user/mo annual, $25 monthly | Fixed, but Standard seats have Plus-level limits sized for interactive use |
| Credit overages | Per-model rates, top models ~50x the light ones | The spiral: bought when devs are blocked, priced at panic rates |
| Raw API keys for agents | Pure metered | None; this is usually the line that doubles |
| Flat routed budget for volume work | $19–$249 per developer key, fixed | Exact, and it absorbs precisely the workload that causes overages |
The split that makes the number fixed
Trying to cover all Codex usage with one billing mode is the mistake. Seats are excellent value for the interactive layer (reviews, quick edits, questions) because a human at a keyboard rarely outruns the window. Agentic volume is the opposite: one long refactor session consumes dozens of requests, and it is exactly this traffic that turns into credit purchases and API spikes.
Route by workload instead: seats for humans, a budgeted endpoint for agents. Codex CLI makes the split trivial: profiles let a dev flip between plan sign-in and the custom provider per invocation, so the same tool bills to the right bucket depending on what it is doing.
Setting it up for a 5-dev team
- Keep Business seats for everyone; the interactive layer stays first-party and fixed.
- Add a flat-rate provider for agentic runs: each dev gets a key with its own plan, config.toml gets a [model_providers] entry, profiles switch between them.
- Kill standing raw API keys for individuals unless a workload genuinely needs them. They are where unbounded spend hides.
- Report one number: seats ($100/mo for five) plus the sum of the approved key plans. Both are fixed, neither can spike.
The catch
The fixed-number setup caps spend, not appetite: if the team's agentic volume outgrows the flat budget, requests stop at the budget or you buy the next tier. That is a visible, planned decision. That beats the alternative, where the same growth shows up as an overage invoice nobody approved.
Where Standard Compute fits
Standard Compute is the budgeted half of the split: Codex CLI points at api.stdcmpt.com/v1 (both /v1/responses and /v1/chat/completions). A free company workspace gives every dev a key with its own plan, from Standard ($89 for a $95 budget) to Pro Plus ($499 for $549), approved by an admin and invoiced to the company, and smart routing sends routine agent turns to efficient models so the budget survives daily use. About a quarter of our active customers run Codex CLI this way.
FAQ
Does ChatGPT Business give more Codex usage than Plus?
Standard seats don't: they get the same per-5-hour estimates as Plus. Premium seats get 5x the Standard allowance and no five-hour limit. Devs who cap out on Plus will cap out on Standard seats too; that workload needs Premium seats, credits, or an external budget.
Can different team members share one Codex plan?
No. Plan allowances are per account, and sharing accounts violates terms. Budgeting for a team happens legitimately at the API layer instead: one endpoint with a key per developer, each with its own fixed budget, which is how Standard Compute's company workspace works.
How much should a 5-dev team budget for AI coding per month?
Real-world band: $100/mo for interactive seats, plus $150–250 per agent-heavy developer if metered, so $500–1,300 total. A budget per key compresses the agentic half to a fixed number: five Standard keys are $445 a month, and heavy agent users move to Pro ($249) or Pro Plus ($499) keys.