Rate Limits for New GPT Image Model

[yangzhi.see]

Hi everyone,

I was looking at the rate limits table (attached at the bottom) and had a couple of questions I was hoping someone could help clarify:

1. Are TPM (Tokens per Minute) and IPM (Images per Minute) separate limits, or are they somehow linked together?
2. Do these limits apply only to input tokens (both text and image), output tokens, or both?

Would appreciate any clarification — thanks in advance!

---

“input_tokens” can consist of the following:

{
      "text_tokens": int(),
      "image_tokens": int()
}

Calculation of image output tokens:

• The number of tokens generated depends on image dimensions and quality:

Quality	Square (1024×1024)	Portrait (1024×1536)	Landscape (1536×1024)
Low	272 tokens	408 tokens	400 tokens
Medium	1056 tokens	1584 tokens	1568 tokens
High	4160 tokens	6240 tokens	6208 tokens

---