{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreighs75b7qcgvmkmb43aosfdli4rdvrvaifzq2tuu2oi24x5qv5yzy",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mkwecvx4jlh2"
  },
  "path": "/t/prompt-for-automation-solution-image/1380179#post_5",
  "publishedAt": "2026-05-03T03:53:24.000Z",
  "site": "https://community.openai.com",
  "textContent": "Here is what gpt-5.5 says about prompting in English:\n\n> For gpt-image-2, English is generally the safest choice, especially for complex, highly specific, or stylistically detailed prompts. Image models are typically trained with a large amount of English captioning and art-direction language, so English prompts may produce more reliable results for nuanced instructions.\n>\n> Your native language can also work well, particularly for simple descriptions or when expressing culturally specific ideas. If the prompt includes idioms, regional references, or subtle visual details, results may vary depending on how well the model understands those terms.\n>\n> In short: English is usually best for maximum consistency, while your native language is acceptable for straightforward prompts or culturally specific concepts.",
  "title": "Prompt for automation solution image"
}