Understanding Limit Details

You can call the Get Limit Details API to see the status of the limit. This page describes the response object and how to interpret what it represents.

{
  "limit_creation_timestamp": "2024-09-01T23:56:55.966655Z",
  "limit_id": "295859",
  "limit_name": "OpenAI quickstart allow limit",
  "limit_type": "allow",
  "limit_update_timestamp": "2024-09-01T23:56:56.440000Z",
  "currency": "usd",
  "max": "12.5",
  "threshold": "0.8",
  "totals": {
    "cost": {
      "input": {
        "base": "0.000007"
      },
      "output": {
        "base": "0.000006"
      },
      "total": {
        "base": "0.000013",
        "overrun_base": "0.0"
      }
    },
    "requests": {
      "ok": 1,
      "exceeded": 0,
      "failed": 0,
      "blocked": 0,
      "blocked_external": 0,
      "total": 1
    }
  },
  "limit_tags": [
    "example_limit"
  ]
}

Cost

The cost object stores the total cost in USD of all of the successful requests that were made with this limit_id specified.

The cost is split between input, output, and total, where total is the sum of input and output.

base and overrun_base

The input and output objects provide the cost of the input and output units, across all Categories and Resources used, for all of the successful requests.

The base cost represents the raw cost of the units as determined by the Resource.

The overrun_base is how much of the base cost exceeded the limit's max, known as the Limit Overrun.

For example, if a limit has:

• max of $10.00
• input base cost of $7.30
• output base cost of $5.21

Then the overrun_base is:

• $7.30 + $5.21 - $10.00 = $2.51

Requests

The requests object provides the number of requests and their status types for which this limit was passed in. Note that only successful requests are considered for calculating the costs against the limit.