{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreichw6caqz7vac2yojrcpjwez2vb6w5cro2uvnw7qbxlwijvxsq6ju",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mpgwn6eneig2"
  },
  "path": "/t/why-is-gpt-5-4-mini-showing-up-in-my-codex-usage/1385210#post_7",
  "publishedAt": "2026-06-29T16:26:49.000Z",
  "site": "https://community.openai.com",
  "tags": [
    "@rajeshamara"
  ],
  "textContent": "VeitB:\n\n> Afaik this is also true for the memory features in Codex.\n\nReally? It would be nice to make that optional.\n\nEspecially considering it’s such a tiny percentage … (if we go by the OP stats) (thanks @rajeshamara )\n\nMy fear is that any errors with small models might put at risk context committed to the bigger models …",
  "title": "Why Is GPT-5.4 Mini Showing Up in My Codex Usage?"
}