Avoiding OpenAI Token Rate Limit

[bsimon11]

Hello, thank you for the exciting work! Do you have any advice for avoiding the OpenAI token rate limit which is set at 30.000 tokens per minute for 4o? I run into this error consistently when using STAgent. Thanks!

[linzw14]

Hi, thanks for using STAgent, and we highly appreciate your feedback!

In our case, we use an organization-level (university) OpenAI API account (Tier 5), which provides a generous rate limit of 30,000,000 tokens per minute. If you're affiliated with a university, consider reaching out to your institution’s IT or research computing department — they may already have a dedicated OpenAI API access.

Alternatively, you can upgrade to Tier 2 (see OpenAI usage tiers). Tier 2 offers 450,000 tokens per minute for gpt-4o, which is typically sufficient for most academic or research use cases.

By contrast, Tier 1 users are limited to 30,000 tokens per minute, which can be a bottleneck when performing spatial omics analysis, especially with large numbers of high-resolution images. Unfortunately, there isn’t a straightforward workaround for this limitation.

Hope this helps clarify your options!