{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreibd3hxq7cw4o4uqejf46fkmrqtuce42gudqyxbhebjdbtewycut7u",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mm2wy7c2mo32"
  },
  "path": "/t/codex-credit-usage-is-inconsistent/1381102#post_2",
  "publishedAt": "2026-05-17T17:29:58.000Z",
  "site": "https://community.openai.com",
  "tags": [
    "support case",
    "@Agentify.sh"
  ],
  "textContent": "Hey @Agentify.sh, this can be confusing.\n\nYou might be already aware of this but charges are based on token usage rather than “per message,” so a single message can cost more or less depending on input size, output size, model, and any additional tool usage. The first thing I’d check is the current rate card and compare it against the usage shown in your dashboard.\n\nIf the numbers still don’t seem to match after that, opening a support case is the right next step. Share the case number here along with the model, approximate dates, and what looks off, and I can help narrow it down.\n\n- Avinash",
  "title": "Codex Credit usage is inconsistent"
}