> ## Documentation Index > Fetch the complete documentation index at: https://docs.kiteml.com/llms.txt > Use this file to discover all available pages before exploring further. # Estimate > What training runs will be charged, in tokens, before you create them. A run on a GPU tier is charged one GPU-hour at the tier's rate, taken when it starts, however long it trains and however it ends; nothing is refunded and nothing more is taken. A run whose GPU job never starts is not charged, and CPU tiers are free. So ``hours`` does not change the answer: it is echoed in ``assumptions``. ## OpenAPI ````yaml /api-reference/openapi.json post /v1/training_runs/estimate openapi: 3.1.0 info: title: Kite Platform API description: >- Public API for Kite. Authenticate with `Authorization: Bearer kite_...`, and send `Kite-Version: 2026-09-27` to pin the dated contract your code was written against (without it you get 2026-09-27). Every response carries the version that served it. version: '2026-09-27' servers: - url: https://api.kiteml.com description: Kite Platform API security: - bearerAuth: [] tags: - name: Augmentations description: Generate new episodes from a LeRobot dataset. - name: Training runs description: Fine-tune a policy on a dataset, one run per policy. - name: Twins description: Rebuild a robot episode as an interactable MuJoCo scene (private beta). - name: Catalog description: Policies, hardware tiers, and dataset inspection. - name: Operations description: Poll any run by its id. - name: Webhooks description: Endpoints that receive events, and the event log. - name: API keys description: The key making the request. - name: Usage description: Token usage and credit balance. paths: /v1/training_runs/estimate: post: tags: - Training runs summary: Estimate description: >- What training runs will be charged, in tokens, before you create them. A run on a GPU tier is charged one GPU-hour at the tier's rate, taken when it starts, however long it trains and however it ends; nothing is refunded and nothing more is taken. A run whose GPU job never starts is not charged, and CPU tiers are free. So ``hours`` does not change the answer: it is echoed in ``assumptions``. operationId: estimate_v1_training_runs_estimate_post parameters: - $ref: '#/components/parameters/KiteVersion' requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/TrainingRunEstimate' responses: '200': description: The estimate content: application/json: schema: $ref: '#/components/schemas/TrainingRunCostEstimate' '400': description: >- Invalid request (`parameter_invalid`, `unsupported_version`, ...); `param` names the field content: application/json: schema: $ref: '#/components/schemas/ApiErrorResponse' '401': description: Missing or invalid API key (`unauthenticated`) content: application/json: schema: $ref: '#/components/schemas/ApiErrorResponse' '403': description: >- The key lacks a scope this route needs (`missing_scope`), or the feature is not enabled for the account content: application/json: schema: $ref: '#/components/schemas/ApiErrorResponse' '429': description: >- Rate or concurrency limited (`rate_limit_exceeded`, `concurrency_limit_exceeded`, `capacity_exceeded`); honor `Retry-After` content: application/json: schema: $ref: '#/components/schemas/ApiErrorResponse' components: parameters: KiteVersion: name: Kite-Version in: header required: false description: >- The dated contract your code was written against. Without it you get 2026-09-27; every response says which version served it. schema: type: string enum: - '2026-09-27' default: '2026-09-27' schemas: TrainingRunEstimate: properties: hardware_tier: type: string maxLength: 64 title: Hardware Tier default: gcp_gpu_t4 policies: type: integer maximum: 8 minimum: 1 title: Policies description: How many runs default: 1 hours: type: number maximum: 72 exclusiveMinimum: 0 title: Hours description: >- How long you expect each run to train. It does not change the charge: a run is charged one GPU-hour however long it trains default: 1 type: object title: TrainingRunEstimate TrainingRunCostEstimate: properties: object: type: string enum: - training_run_estimate const: training_run_estimate title: Object tokens: type: integer title: Tokens description: >- What these runs are charged together, `tokens_per_run` × `assumptions.policies`. Your balance must cover it to create them tokens_per_run: type: integer title: Tokens Per Run description: >- What one run is charged: one GPU-hour at the tier's `tokens_per_hour`, however long it trains, taken when it starts. A run whose GPU job never starts is not charged. 0 on a CPU or unknown tier assumptions: $ref: '#/components/schemas/TrainingRunEstimateAssumptions' description: The request, with its defaults filled in type: object required: - object - tokens - tokens_per_run - assumptions title: TrainingRunCostEstimate description: What training would cost, in tokens. examples: - assumptions: hardware_tier: gcp_gpu_t4 hours: 1 policies: 2 object: training_run_estimate tokens: 1000 tokens_per_run: 500 ApiErrorResponse: properties: error: $ref: '#/components/schemas/ApiErrorBody' type: object required: - error title: ApiErrorResponse description: Every /v1 error has this shape. TrainingRunEstimateAssumptions: properties: hardware_tier: type: string title: Hardware Tier policies: type: integer title: Policies description: How many runs hours: type: number title: Hours description: From the request. It does not change the charge type: object required: - hardware_tier - policies - hours title: TrainingRunEstimateAssumptions ApiErrorBody: properties: type: type: string title: Type description: >- invalid_request_error, authentication_error, permission_error, not_found_error, conflict_error, rate_limit_error, or api_error code: type: string title: Code description: 'Machine-readable: branch on this, not on the message' message: type: string title: Message param: anyOf: - type: string - type: 'null' title: Param description: The request field the error is about, when there is one request_id: anyOf: - type: string - type: 'null' title: Request Id description: Also in the X-Request-Id header; quote it to support details: anyOf: - type: object - type: 'null' title: Details description: >- Extra data for some codes (e.g. the cameras a dataset does have, for camera_not_found) type: object required: - type - code - message title: ApiErrorBody securitySchemes: bearerAuth: type: http scheme: bearer description: >- A Kite API key (`kite_...`), from **Platform API → Keys** in the dashboard. ````