Limits
The units, requests, tokens, sizes, and times AI allows.
Your plan
| Limit | Free |
|---|---|
| AI units in a month | 100,000 |
| Requests in a minute | 20 |
| Tokens in a minute | 40,000 |
Tokens in a minute count each call’s input and its max_tokens of output, when it is sent. A project’s Limits page in the console shows its plan’s limits, and you can set a lower monthly limit of your own. See Track usage and set limits.
Calls
| Limit | Value |
|---|---|
| Chat request body | 8 MB |
| Management request body | 1 MB |
| Messages in one call | 1,000 |
| Stop sequences | 4 |
| Choices in one answer | 1 |
user | 256 characters |
| Time a call can take | 10 minutes |
| Keep-alive while streaming | Every 15 seconds |
Each model’s context and most output are in Models.
Prompts
| Limit | Value |
|---|---|
| Prompts in one project | 100 |
| Instructions | 32 KB |
| Variables in one prompt | 50 |
| A variable’s value | 32 KB |
| Time for a change to reach calls without a pinned version | Up to 1 minute |
Records
| Limit | Value |
|---|---|
| Record of each call | Kept 90 days, without prompts or answers |
One MB is 1,048,576 bytes.
Last updated