{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigfnokdcc2ceqosrgufraie4udsk7n2dop4ejdc2ouoofraqmllyi",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mm35oijg62h2"
  },
  "path": "/t/codex-using-credits-and-charging-undue-payments/1380373#post_2",
  "publishedAt": "2026-05-17T19:33:29.000Z",
  "site": "https://community.openai.com",
  "tags": [
    "@devstone"
  ],
  "textContent": "Hey @devstone, yeah, I can see why that would be confusing, especially if it looked like credits were being used while you weren’t actively sending prompts.\n\nOne thing to keep in mind is that Codex/API costs are tied to actual usage, like tokens, model choice, context size, outputs, and any tools or tasks running, not just the number of messages you send. I’d first compare the usage shown in your dashboard against the current rate card and the timing of the activity.\n\nIf the numbers still don’t line up, please open a support case with support@openai.com and include the approximate dates/times, model used, screenshots if you have them, and what specifically looks off. If you already have a case number, share it here and I can take a look.\n\n- Sunny",
  "title": "Codex using credits and charging undue payments"
}