{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreig3jdie2seb4rztetnrnbenaj2dt5fjejsnk6ujd76g6ltbyafgp4",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mmcc3kbthjg2"
  },
  "path": "/t/why-do-gpt-5-1-and-gpt-5-4-mini-behave-so-differently-in-production-chatbot-use-cases/1380891#post_12",
  "publishedAt": "2026-05-20T16:08:16.000Z",
  "site": "https://community.openai.com",
  "textContent": "I find that both are outdated and buggy as they were originally released.\n\nI know that with codex I spend more time fixing the limited effects done by 5.4 and if just isn’t worth the pain. As for ChatGPT I don’t use it enough other than I’ll pay the Piper and stick with the better models.",
  "title": "Why do gpt-5.1 and gpt-5.4-mini behave so differently in production chatbot use cases?"
}