Rate limits (Google add-on only)
GPT for Sheets and Docs is built on OpenAI and Google services. Both services enforce rate limits in order to ensure a smooth and fair access to everyone.
OpenAI rate limits
The table below provides guidance on the rate limits typically enforced by OpenAI, based on the type of plan and the model used.
| ChatGPT model (gpt-3.5-turbo) | GPT-3 base models | GPT-4 | GPT-4-32k | |
|---|---|---|---|---|
| Free trial users | 20 requests / minute 150,000 tokens / minute | 20 requests / minute 150,000 tokens / minute | ||
| Pay-as-you-go users (in your first 48 hours) | 60 requests / minute 60,000 tokens / minute | 60 requests / minute 250,000 tokens / minute | ||
| Pay-as-you-go users (after your first 48 hours) | 3,500 requests / minute 90,000 davinci tokens / minute (and proportionally more for smaller models) | 3,500 requests / minute 350,000 davinci tokens / minute (and proportionally more for smaller models) | 200 requests / minute 40,000 tokens / minute | 400 requests / minute 80,000 tokens / minute |
Google rate limits
The table below provides guidance on the rate limits typically enforced by Google, based on the type of account.
| External service calls daily limits | |
|---|---|
| gmail.com | 20k / day / gmail.com account |
| Google Workspace | 100k / day / Google Workspace account |