Get Account Usage ​

Returns personal account Usage, Provider, and Performance analytics with the same query granularity as Analysis > Usage in the 开桥 console.

GET https://openqiao.com/api/v1/management/usage

Authentication ​

Authorization Header * ​

Authorization: Bearer <OPENQIAO_MANAGEMENT_API_KEY>

• Name: Authorization
• Format: Bearer <API_KEY>
• Description: A Management API Key created using your personal account in the 开桥 Console

Query parameters ​

For type=usage or type=performance, omitting api_key_ids, model_slugs, and bill_types (or leaving them empty) returns all summary and grouped statistics for the personal account that owns the Management Key, within the period specified by query_dimension and query_time.

For type=provider, model_slugs is still required and must contain exactly one model. api_key_ids and bill_types can be omitted.

BIZ_MTH returns the selected month's daily series, BIZ_DT returns the selected day's hourly series, and BIZ_HOUR returns the selected hour's minute series. For BIZ_HOUR, convert the selected local hour to UTC before creating query_time, as the console does.

Whitespace and duplicate CSV values are removed. Unknown query parameters and invalid calendar values return HTTP 400.

api_key_ids only accepts IDs of non-deleted inference or subscription API Keys belonging to the authenticated personal account. Passing a Management Key ID, another account's Key ID, or a deleted Key ID returns HTTP 403.

Returns ​

success boolean ​

Whether the request succeeded. Always true in a successful response.

data object ​

Query result whose shape depends on type. All totals and grouped data below use the account, time range, and filters from this request.

Non-null amounts, token counts, counts, ratios, and performance metrics are decimal strings, not JSON numbers. null means no statistic is available or the field is not used for that grouping; it is not equivalent to "0". A successful response may contain empty arrays and null summary fields when no data is available.

Time series use the following bizTime formats according to query_dimension:

Usage (type=usage) ​

data.tokensTotal string | null ​

Total tokens for the selected filters.

data.tokensPrompt string | null ​

Input token count for the selected filters.

data.tokensCompletion string | null ​

Output token count for the selected filters.

data.requestCounts string | null ​

Total requests for the selected filters.

tokensByModel, tokensByTokenType, and tokensByApiKey are different groupings of the same usage; do not add token counts across these arrays.

data.tokensByModel array ​

Grouped by time and model. Each item contains:

• bizTime string — Time bucket identifier in the format listed above, not an individual request timestamp.
• modelSlug string — Model identifier, such as openai/gpt-5, not its display name.
• tokenType string | null — Not a token-type grouping; normally null.
• apiKeyId string | null — Not an API Key grouping; normally null.
• tokens string | null — Token count for this group and time bucket.
• requestCounts string | null — Request count for this group and time bucket.

data.tokensByTokenType array ​

Grouped by time and token type. Each item contains:

• bizTime string — Time bucket identifier in the format listed above, not an individual request timestamp.
• modelSlug string | null — Not a model grouping; normally null.
• tokenType string — Billable token type, such as prompt (input) or completion (output).
• apiKeyId string | null — Not an API Key grouping; normally null.
• tokens string | null — Token count for this group and time bucket.
• requestCounts string | null — Request counts are normally not returned for this grouping and are null.

data.tokensByApiKey array ​

Grouped by time and API Key. Each item contains:

• bizTime string — Time bucket identifier in the format listed above, not an individual request timestamp.
• modelSlug string | null — Not a model grouping; normally null.
• tokenType string | null — Not a token-type grouping; normally null.
• apiKeyId string — API Key resource ID accepted by api_key_ids, not the secret key token.
• tokens string | null — Token count for this group and time bucket.
• requestCounts string | null — Request count for this group and time bucket.

data.countsByWebSearch array ​

Web Search requests grouped by time and model. Each item contains:

• bizTime string — Time bucket identifier in the format listed above, not an individual request timestamp.
• modelSlug string — Model identifier, such as openai/gpt-5, not its display name.
• tokenType string | null — Not a token-type grouping; normally null.
• apiKeyId string | null — Not an API Key grouping; normally null.
• tokens string | null — Web Search is counted in requests; normally null, not a token-usage measurement for searches.
• requestCounts string | null — Web Search request count for this model and time bucket; do not add it to the overall request total.

Provider (type=provider) ​

model_slugs requires exactly one model. data returns endpoint statistics for that model within the selected filters:

data.fastestEndpointSlug string | null ​

Identifier of the fastest endpoint within the query; may be null when unavailable.

data.mainEndpointSlug string | null ​

Identifier of the main endpoint with the most requests within the query; may be null when unavailable.

data.uptime string | null ​

Overall request success rate for the selected model, from 0 to 1. For example, "0.9920" means 99.20%; this is not elapsed uptime.

data.endpointCount string | null ​

Number of endpoints within the query, represented as a string.

data.endpointUsages array ​

Statistics grouped by endpoint. Each item contains:

• endpointSlug string — Endpoint identifier distinguishing the endpoints used for the selected model.
• requestCounts string | null — Total requests for this endpoint.
• tokens string | null — Total tokens for this endpoint.
• requestRatio string | null — Share of requests served by this endpoint within the query, from 0 to 1; "0.8" means 80%.
• tokenRatio string | null — Share of tokens served by this endpoint within the query, from 0 to 1.
• medianLatency string | null — Median request latency for this endpoint, in milliseconds (ms), not mean latency.
• medianThroughput string | null — Median throughput for this endpoint, in tokens per second.
• billAmount string | null — Cost for this endpoint, in USD.
• billAmountRatio string | null — Share of the query’s total cost attributable to this endpoint, from 0 to 1.

Performance (type=performance) ​

data.medianLatency string | null ​

Median request latency for the selected filters, in milliseconds (ms), not mean latency.

data.medianThroughput string | null ​

Median throughput for the selected filters, in tokens per second.

Do not compute overall medians by averaging model-level medians.

data.modelPerformances array ​

Performance summary grouped by model. Each item contains:

• modelSlug string — Model identifier, such as openai/gpt-5, not its display name.
• medianLatency string | null — Median request latency for this model, in milliseconds (ms).
• medianThroughput string | null — Median throughput for this model, in tokens per second.

data.latencyDetails array ​

Latency time series grouped by time and model. Each item contains:

• bizTime string — Time bucket identifier in the format listed above, not an individual request timestamp.
• modelSlug string — Model identifier, such as openai/gpt-5, not its display name.
• medianLatency string | null — Median request latency for this model and time bucket, in milliseconds (ms).

data.throughputDetails array ​

Throughput time series grouped by time and model. Each item contains:

• bizTime string — Time bucket identifier in the format listed above, not an individual request timestamp.
• modelSlug string — Model identifier, such as openai/gpt-5, not its display name.
• medianThroughput string | null — Median throughput for this model and time bucket, in tokens per second.

curl -G https://openqiao.com/api/v1/management/usage \
  -H "Authorization: Bearer $OPENQIAO_MANAGEMENT_API_KEY" \
  --data-urlencode "type=usage" \
  --data-urlencode "query_dimension=BIZ_MTH" \
  --data-urlencode "query_time=202609" \
  --data-urlencode "api_key_ids=key_1,key_2" \
  --data-urlencode "model_slugs=openai/gpt-5" \
  --data-urlencode "bill_types=metered,subscription"

{
  "success": true,
  "data": {
    "tokensTotal": "1500000",
    "tokensPrompt": "1000000",
    "tokensCompletion": "500000",
    "requestCounts": "125",
    "tokensByModel": [
      {
        "bizTime": "20260903",
        "modelSlug": "openai/gpt-5",
        "tokenType": null,
        "apiKeyId": null,
        "tokens": "1500000",
        "requestCounts": "125"
      }
    ],
    "tokensByTokenType": [
      {
        "bizTime": "20260903",
        "modelSlug": null,
        "tokenType": "prompt",
        "apiKeyId": null,
        "tokens": "1000000",
        "requestCounts": null
      },
      {
        "bizTime": "20260903",
        "modelSlug": null,
        "tokenType": "completion",
        "apiKeyId": null,
        "tokens": "500000",
        "requestCounts": null
      }
    ],
    "tokensByApiKey": [
      {
        "bizTime": "20260903",
        "modelSlug": null,
        "tokenType": null,
        "apiKeyId": "key_1",
        "tokens": "1000000",
        "requestCounts": "80"
      },
      {
        "bizTime": "20260903",
        "modelSlug": null,
        "tokenType": null,
        "apiKeyId": "key_2",
        "tokens": "500000",
        "requestCounts": "45"
      }
    ],
    "countsByWebSearch": [
      {
        "bizTime": "20260903",
        "modelSlug": "openai/gpt-5",
        "tokenType": null,
        "apiKeyId": null,
        "tokens": null,
        "requestCounts": "3"
      }
    ]
  }
}

curl -G https://openqiao.com/api/v1/management/usage \
  -H "Authorization: Bearer $OPENQIAO_MANAGEMENT_API_KEY" \
  --data-urlencode "type=provider" \
  --data-urlencode "query_dimension=BIZ_DT" \
  --data-urlencode "query_time=20260903" \
  --data-urlencode "model_slugs=openai/gpt-5"

{
  "success": true,
  "data": {
    "fastestEndpointSlug": "provider-a",
    "mainEndpointSlug": "provider-a",
    "uptime": "0.9920",
    "endpointCount": "2",
    "endpointUsages": [
      {
        "endpointSlug": "provider-a",
        "requestCounts": "100",
        "tokens": "1200000",
        "requestRatio": "0.8000",
        "tokenRatio": "0.8000",
        "medianLatency": "320",
        "medianThroughput": "85.5",
        "billAmount": "9.1234000000",
        "billAmountRatio": "0.7390"
      },
      {
        "endpointSlug": "provider-b",
        "requestCounts": "25",
        "tokens": "300000",
        "requestRatio": "0.2000",
        "tokenRatio": "0.2000",
        "medianLatency": "410",
        "medianThroughput": "76.5",
        "billAmount": "3.2222000000",
        "billAmountRatio": "0.2610"
      }
    ]
  }
}

curl -G https://openqiao.com/api/v1/management/usage \
  -H "Authorization: Bearer $OPENQIAO_MANAGEMENT_API_KEY" \
  --data-urlencode "type=performance" \
  --data-urlencode "query_dimension=BIZ_HOUR" \
  --data-urlencode "query_time=2026090312" \
  --data-urlencode "model_slugs=openai/gpt-5,anthropic/claude-sonnet-4"

{
  "success": true,
  "data": {
    "medianThroughput": "82.5",
    "medianLatency": "340",
    "modelPerformances": [
      {
        "modelSlug": "openai/gpt-5",
        "medianLatency": "320",
        "medianThroughput": "85.5"
      },
      {
        "modelSlug": "anthropic/claude-sonnet-4",
        "medianLatency": "360",
        "medianThroughput": "79.5"
      }
    ],
    "latencyDetails": [
      {
        "bizTime": "202609031215",
        "modelSlug": "openai/gpt-5",
        "medianLatency": "320"
      },
      {
        "bizTime": "202609031215",
        "modelSlug": "anthropic/claude-sonnet-4",
        "medianLatency": "360"
      }
    ],
    "throughputDetails": [
      {
        "bizTime": "202609031215",
        "modelSlug": "openai/gpt-5",
        "medianThroughput": "85.5"
      },
      {
        "bizTime": "202609031215",
        "modelSlug": "anthropic/claude-sonnet-4",
        "medianThroughput": "79.5"
      }
    ]
  }
}

Rate limits and errors ​

Account Usage and Cost endpoints share a limit of 60 requests per minute for each personal account. Request 61 returns HTTP 429 and Retry-After: 60. Existing platform-wide Management API limits also apply.