{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreicpjhl2yiv2d6nv5r4gnvulqnyi4pa6em7hxx6mnoux42uemm2y5e",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mn6msy4h3i62"
  },
  "path": "/t/need-to-have-serious-discussion-with-the-200-month-plan/1382162#post_2",
  "publishedAt": "2026-05-31T21:58:00.000Z",
  "site": "https://community.openai.com",
  "textContent": "Yes, I was also one of the people who felt that there had been a regression in that regard. However, at this point, I don’t think the decline is nearly as severe as some people suggest. There are clear signs of recovery.\n\nAlso, when GPT-5.6 is released soon, I think switching to Opus 4.8 might be something you regret, because OpenAI has consistently been very strong at pushing the frontier forward and releasing models that stay one step ahead.\n\nTo me, Opus 4.8 feels like a balancing model somewhere between Gemini and GPT, taking certain strengths from both. Gemini tends to be highly creative and energetic, but it also hallucinates more and can be overly enthusiastic. Its tool usage and context tracking are not as reliable. ChatGPT, on the other hand, tends to be less imaginative and more conservative, almost the opposite of Gemini in that sense.\n\nClaude models feel like they sit somewhere in the middle, trying to balance those two extremes.\n\nHowever, when it comes to heavy reasoning, serious coding, and complex architectural work, I currently don’t think anything beats GPT models. If you are building a large, complex architecture, moving away from GPT would likely be a serious strategic mistake.",
  "title": "Need to have serious discussion with the $200/month plan"
}