{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreifqjky6juiashes4z7pmotgaacszri5wuf57dcbqd34qtaabzoooi",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mgd3nwd3agt2"
  },
  "path": "/t/question-about-image-rate-limits-with-gpt-5-3-chat-when-using-previous-response-id/1375787#post_1",
  "publishedAt": "2026-03-05T14:45:15.000Z",
  "site": "https://community.openai.com",
  "textContent": "Hi everyone,\n\nI ran into an image rate limit while testing `gpt-5.3-chat-latest` during a Twitch stream integration, and I’m trying to understand exactly how the **input image per minute limit** works when using `previous_response_id`.\n\n### Setup\n\nI run a Python middleware that connects Twitch chat and game screenshots to the OpenAI Responses API. The bot has two main interaction types:\n\n  1. **Text responses**\n\n     * Twitch chat → LLM response\n\n     * No images included\n\n  2. **Screenshot reactions**\n\n     * A screenshot from the game is sent with a prompt\n\n     * The model comments on what it sees\n\n\n\n\nBoth routes use `previous_response_id` so the model maintains conversational context.\n\n### What I observed\n\nWhen using `gpt-5.3-chat-latest`, I started hitting this error:\n\n> Rate limit reached for input-images per min\n>  Limit: 10\n>  Used: 4\n>  Requested: 7\n\nThe confusing part is that **the request that triggered the error did not include any images**. It was a text-only message.\n\nHowever, earlier in the conversation thread there had been several screenshot reactions.\n\n### Behavior difference between models\n\nDuring the same stream I switched the model to:\n\n`gpt-5.1-chat-latest`\n\nAfter switching, I sent **many more screenshots over time** and never hit the same image rate limit.\n\nThis makes me suspect one of the following:\n\n  1. `gpt-5.3-chat` has a **lower image-per-minute limit** than `gpt-5.1`\n\n  2. When using `previous_response_id`, **earlier image turns may be reprocessed as part of the context window** , causing a later text-only request to count as multiple image inputs\n\n  3. Some other internal behavior specific to `gpt-5.3`’s multimodal context handling\n\n\n\n\n### Questions\n\nI’m hoping someone from OpenAI or another developer familiar with this can clarify:\n\n  1. Are **input image limits different per model** , specifically between `gpt-5.1` and `gpt-5.3`?\n\n  2. When using `previous_response_id`, can **earlier images in the thread count again toward the image/minute limit** if they are included in context?\n\n  3. Is the **10 images/min limit expected for this model** , and is it temporary or likely to change?\n\n  4. For systems that mix **text chat and occasional screenshots** , is the recommended approach to:\n\n     * separate text and image interactions into **different conversation threads** , or\n\n     * avoid continuing screenshot threads with `previous_response_id`?\n\n\n\n\n### Context for usage\n\nThis system runs during a live stream, so stability is important. A screenshot may be sent every 90–120 seconds, but normal chat messages continue in between.\n\nUnderstanding how the image rate limit interacts with conversation threading would help determine the best architecture.\n\nThanks!",
  "title": "Question about image rate limits with gpt-5.3-chat when using previous_response_id"
}