> ## Documentation Index > Fetch the complete documentation index at: https://docs.galtea.ai/llms.txt > Use this file to discover all available pages before exploring further. # Get the estimated cost, LLM spans and tokens of a period split by model, version or end user > The sessions `GET /analytics/cost-overview` counts, split into one row per model, version or end user, with every row returned. The rows add up to the `totalCost` of `GET /analytics/cost-overview` for the same filters. The end user is only set on turns built from OpenTelemetry spans. Needs `evaluation.readInference` on every product of the sum, else 403. ## OpenAPI ````yaml /openapi.json get /analytics/cost-breakdown openapi: 3.0.0 info: version: 1.0.0 title: Product Management Service API description: API documentation for Product Management Service contact: name: Galtea AI servers: - url: https://api.galtea.ai security: - bearerAuth: [] tags: [] externalDocs: description: Galtea Platform Documentation url: https://docs.galtea.ai paths: /analytics/cost-breakdown: get: tags: - analytics summary: >- Get the estimated cost, LLM spans and tokens of a period split by model, version or end user description: >- The sessions `GET /analytics/cost-overview` counts, split into one row per model, version or end user, with every row returned. The rows add up to the `totalCost` of `GET /analytics/cost-overview` for the same filters. The end user is only set on turns built from OpenTelemetry spans. Needs `evaluation.readInference` on every product of the sum, else 403. operationId: getCostBreakdown parameters: - schema: type: string format: date-time description: Start of the period, inclusive (ISO 8601) required: true description: Start of the period, inclusive (ISO 8601) name: from in: query - schema: type: string format: date-time description: >- End of the period, exclusive (ISO 8601). A period that does not end after `from`, or is longer than 1830 days (5 years), is rejected with 400. required: true description: >- End of the period, exclusive (ISO 8601). A period that does not end after `from`, or is longer than 1830 days (5 years), is rejected with 400. name: to in: query - schema: type: boolean description: >- Count production sessions when true, development sessions when false. required: true description: >- Count production sessions when true, development sessions when false. name: isProduction in: query - schema: type: string description: >- The organization whose products to read. Defaults to the caller's own organization. Another organization reads products only for a caller that may read across organizations. required: false description: >- The organization whose products to read. Defaults to the caller's own organization. Another organization reads products only for a caller that may read across organizations. name: organizationId in: query - schema: type: string description: Keep the products whose name contains this (case-insensitive) required: false description: Keep the products whose name contains this (case-insensitive) name: name in: query - schema: type: array items: type: string description: Keep the products that carry any of these tags required: false description: Keep the products that carry any of these tags style: form explode: true name: tags in: query - schema: type: array items: type: string description: >- Keep only these products. A product outside the caller's grant is refused with 403; a product of another organization, or a deleted one, matches nothing. At most 25. maxItems: 25 required: false description: >- Keep only these products. A product outside the caller's grant is refused with 403; a product of another organization, or a deleted one, matches nothing. At most 25. style: form explode: true name: productIds in: query - schema: type: string enum: - model - version - endUser description: >- The value each row stands for: the model that answered each LLM span, the version, or the end user of each turn example: model required: true description: >- The value each row stands for: the model that answered each LLM span, the version, or the end user of each turn name: groupBy in: query responses: '200': description: The cost of the products per value content: application/json: schema: $ref: '#/components/schemas/CostBreakdown' '400': description: Bad request content: application/json: schema: $ref: '#/components/schemas/Error' '401': description: Unauthorized content: application/json: schema: $ref: '#/components/schemas/Error' '403': description: Forbidden content: application/json: schema: $ref: '#/components/schemas/Error' security: - bearerAuth: [] components: schemas: CostBreakdown: type: object properties: rows: type: array items: $ref: '#/components/schemas/CostBreakdownRow' description: >- Every value, the highest cost first. Each turn's estimated cost is split over the LLM spans it counts, each LLM call once and under the model it reports; the part of a turn's cost that its spans do not carry goes to the version model. So the rows add up to `total`, which is the `totalCost` of `GET /analytics/cost-overview` for the same filters. total: type: object properties: llmSpanCount: type: integer description: >- Spans the turns count for their usage, one each, plus one for each turn that counts none but reports a cost or tokens itself. A call whose input and output sit on two spans counts as two. inputTokens: type: integer description: Input tokens of those LLM spans outputTokens: type: integer description: Output tokens of those LLM spans cost: type: number nullable: true description: Estimated cost. Null when nothing counted carries a cost. required: - llmSpanCount - inputTokens - outputTokens - cost description: The sum of every row required: - rows - total Error: type: object properties: error: type: string example: Error type message: type: string example: Error message description CostBreakdownRow: type: object properties: key: type: string nullable: true description: >- The grouped value: the model name in lowercase, the version id, or the end user. Null when neither the span nor its version names a model, or when the turn has no end user. label: type: string nullable: true description: >- The model name, the version label (v, or v: when the version has a custom name), or the end user llmSpanCount: type: integer description: >- Spans the turns count for their usage, one each, plus one for each turn that counts none but reports a cost or tokens itself. A call whose input and output sit on two spans counts as two. inputTokens: type: integer description: Input tokens of those LLM spans outputTokens: type: integer description: Output tokens of those LLM spans cost: type: number nullable: true description: Estimated cost. Null when nothing counted carries a cost. required: - key - label - llmSpanCount - inputTokens - outputTokens - cost securitySchemes: bearerAuth: type: http scheme: bearer description: >- API key authorization. Pass your API key in the Authorization header as a Bearer token. Both new (`gsk_*`) and legacy (`gsk-`) API keys are accepted, e.g. `Authorization: Bearer gsk_...` or `Authorization: Bearer gsk-...`. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.