Finout has launched the first native FinOps integration for OpenAI Codex Enterprise. It converts Codex's credit-based billing into dollar costs and allocates that spend by team, product, and cost center, the same way Finout already handles cloud, Kubernetes, and every other AI provider it tracks.
At a Glance
Finout is the first FinOps platform to natively convert OpenAI Codex Enterprise's credit-based billing into a dollar figure finance and engineering can both work with. Until now, Codex spend lived in its own credit ledger, disconnected from every other cost a company tracks, which meant a separate report to build and reconcile by hand every billing cycle.
The timing lines up with a real shift in how Codex is used inside companies. Coding agents are moving from experiment to default workflow, and Codex adoption reflects that: among active enterprise users, it rose from near 0% in August 2025 to about 17% by June 2026, per OpenAI's own research, first reported by Axios. As that adoption accelerates, Codex's credit-based billing, split across models and reasoning-effort tiers, becomes harder to attribute to a specific team or reconcile against a budget.
Once connected, Codex credits are converted into dollars using each customer's own contract rate, then sliced by user, model, and reasoning effort. A new OpenAI Origin dimension separates Codex spend from OpenAI Platform spend inside the same account, so the two don't get blended together.
Finout's Virtual Tags- the rules engine for mapping raw spend to business context, map Codex spend to teams, products, and cost centers alongside OpenAI Platform, Anthropic, Cursor, and cloud costs inside MegaBill, Finout's unified billing layer. The result: finance and engineering work from one bill and one anomaly feed, instead of a separate Codex report someone has to reconcile by hand.
"Connecting OpenAI Codex inside Finout finally let us allocate AI spend by team and improve our ROI projections instead of guessing at them."
— Head of FinOps, large B2B SaaS company
The integration extends Finout's Anomaly Detection to Codex usage. A common cost spike is a team shifting from medium to high reasoning effort, which can multiply per-task cost overnight. Finout flags that shift as it happens, rather than after it has already compounded across a full billing cycle.
Codex spend also flows into Billy, Finout's AI FinOps assistant, so teams can ask plain-English questions, like which team drove this week's AI spend spike, and get a live, chart-backed answer without building a dashboard or writing a query.
Connecting the integration takes two inputs: a Codex Enterprise API key and a workspace ID. Add both inside Finout, and usage starts flowing in from the Codex Analytics API on a daily pull, backfilled three months on connection.
There's no proxy sitting in the request path and no separate collector to deploy. The usage data comes straight from Codex's own Analytics API, the same one already running behind your Codex Enterprise account.
The integration is available now to Finout customers running Codex on ChatGPT Enterprise.