{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiewsgfa2lz522u72fuoanp7ymphqevvsyaxrbdiuhlss665wv3eae",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mke2ct2blnq2"
  },
  "path": "/t/collection-of-gpt-image-generator-2-0-issues-bugs-and-work-around-tips-check-first-post/1379535?page=4#post_61",
  "publishedAt": "2026-04-25T21:20:13.000Z",
  "site": "https://community.openai.com",
  "textContent": "Here’s two images one from API and one from ChatGPT. The prompt I used `Generate an image of sunset in Mediterranean.`\n\nAPI (gpt-image-2)\n\nChatGPT\n\nSo in this case no real difference. The thing is, that sometimes it’s unclear which image generator is under the hood in ChatGPT (that’s what I’ve heard before, don’t know about now..so I’m not going to speculate about it).\n\nAlso in API, while prompting images you have access to settings as well as advanced settings. Now in this case I didn’t use anything specific.\n\nAlso in ChatGPT if the memory is on, it usually affects outputs in the same session, meaning even with a new prompt the new image can have resembling the older images on the same session.\n\nDon’t know if this makes anything any clearer, but this is my experiences with API and ChatGPT.",
  "title": "Collection of GPT-image-generator 2.0 issues, bugs, and work-around tips (check first post)"
}