Prime Intellect Training API

openapi: 3.1.0
info:
  title: Prime Intellect Training API
  version: 0.1.0
  description: Hosted reinforcement learning post-training service. Launch dedicated training runs against base models (Qwen3.5,
    Llama, Nemotron, gpt-oss), retrieve token-level usage and cost for each run, and inspect wallet balance. Underpins the
    Prime Intellect Lab managed training workflow.
  contact:
    name: Prime Intellect
    url: https://www.primeintellect.ai
servers:
- url: https://api.primeintellect.ai
security:
- HTTPBearer: []
tags:
- name: training
- name: Billing
paths:
  /api/v1/billing/runs/{run_id}/usage:
    get:
      tags:
      - Billing
      summary: Token usage and cost for a single RFT run
      description: 'Return the running token + cost totals for one RFT run.


        Mirrors the per-row training data shown on the billing page (Training

        tokens, Inference input/output tokens, price per million tokens, and

        total cost) so an agent can poll this endpoint to monitor a live run.'
      operationId: get_run_usage_api_v1_billing_runs__run_id__usage_get
      security:
      - HTTPBearer: []
      parameters:
      - name: run_id
        in: path
        required: true
        schema:
          type: string
          title: Run Id
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/RunUsageResponse'
        '401':
          description: Authorization failed
        '422':
          description: Invalid request data
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
  /api/v1/billing/wallet:
    get:
      tags:
      - Billing
      summary: Wallet balance + most recent billing rows
      description: 'Return the wallet''s current balance and most recent billing rows.


        All resource types are included (compute, training, inference, disks,

        sandboxes, images) — same source-of-truth `Billing` table the

        dashboard''s Billing History tab reads. Sorted by `lastBilledAt` desc.'
      operationId: get_wallet_api_v1_billing_wallet_get
      security:
      - HTTPBearer: []
      parameters:
      - name: limit
        in: query
        required: false
        schema:
          type: integer
          maximum: 100
          minimum: 1
          description: Number of recent billing rows to include (max 100).
          default: 20
          title: Limit
        description: Number of recent billing rows to include (max 100).
      - name: offset
        in: query
        required: false
        schema:
          type: integer
          minimum: 0
          description: Skip this many rows before returning. Lets callers page.
          default: 0
          title: Offset
        description: Skip this many rows before returning. Lets callers page.
      - name: teamId
        in: query
        required: false
        schema:
          anyOf:
          - type: string
          - type: 'null'
          description: If provided, return the team's wallet (requires membership).
          title: Teamid
        description: If provided, return the team's wallet (requires membership).
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/WalletResponse'
        '401':
          description: Authorization failed
        '422':
          description: Invalid request data
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
  /api/v1/training/runs:
    post:
      tags:
      - training
      summary: Create Dedicated Run
      description: Dispatch a dedicated full-FT prime-rl run on a registered PrimeCluster.
      operationId: create_dedicated_run_api_v1_training_runs_post
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CreateDedicatedRunRequest'
        required: true
      responses:
        '201':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/CreateDedicatedRunResponse'
        '401':
          description: Authorization failed
        '422':
          description: Invalid request data
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
      security:
      - HTTPBearer: []
  /api/v1/training/runs/{run_id}:
    delete:
      tags:
      - training
      summary: Delete Dedicated Run
      description: 'Delete a dedicated run: helm uninstall + namespace delete + DB cleanup.


        Idempotent. Safe to call repeatedly. Cleans up the linked Job row too

        (which uninstalls the helm release on the customer cluster). Soft-

        deletes the RFTRun row (matches the existing run-delete UX — Billing

        refs are preserved).'
      operationId: delete_dedicated_run_api_v1_training_runs__run_id__delete
      security:
      - HTTPBearer: []
      parameters:
      - name: run_id
        in: path
        required: true
        schema:
          type: string
          title: Run Id
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/DeleteRunResponse'
        '401':
          description: Authorization failed
        '422':
          description: Invalid request data
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ErrorResponse'
components:
  securitySchemes:
    HTTPBearer:
      type: http
      scheme: bearer
  schemas:
    ErrorResponse:
      properties:
        errors:
          items:
            $ref: '#/components/schemas/ErrorDetail'
          type: array
          title: Errors
      type: object
      required:
      - errors
      title: ErrorResponse
    ErrorDetail:
      properties:
        param:
          type: string
          title: Param
        details:
          type: string
          title: Details
      type: object
      required:
      - param
      - details
      title: ErrorDetail
    WalletResponse:
      properties:
        wallet_id:
          type: string
          title: Wallet Id
        team_id:
          anyOf:
          - type: string
          - type: 'null'
          title: Team Id
        balance_usd:
          type: number
          title: Balance Usd
          description: Current balance in USD
        currency:
          type: string
          title: Currency
        total_billings:
          type: integer
          title: Total Billings
          description: Total billing rows on the wallet (across all pages)
        recent_billings:
          items:
            $ref: '#/components/schemas/BillingEntry'
          type: array
          title: Recent Billings
      type: object
      required:
      - wallet_id
      - balance_usd
      - currency
      - total_billings
      - recent_billings
      title: WalletResponse
      description: 'Snapshot of a wallet''s balance + most recent billing rows.


        Drives `prime wallet`. One call returns everything an agent needs to

        audit the billing flow against a known training run.'
    BillingEntry:
      properties:
        id:
          type: string
          title: Id
        created_at:
          type: string
          format: date-time
          title: Created At
        updated_at:
          type: string
          format: date-time
          title: Updated At
        last_billed_at:
          anyOf:
          - type: string
            format: date-time
          - type: 'null'
          title: Last Billed At
        amount_usd:
          type: number
          title: Amount Usd
        currency:
          type: string
          title: Currency
        resource_type:
          type: string
          title: Resource Type
        resource_id:
          anyOf:
          - type: string
          - type: 'null'
          title: Resource Id
      type: object
      required:
      - id
      - created_at
      - updated_at
      - amount_usd
      - currency
      - resource_type
      title: BillingEntry
      description: 'A single row from the Billing table, with the resource derived from

        whichever foreign-key column is non-null.


        `resource_type` mirrors the strings the dashboard''s billing page filters

        by (`compute`, `training`, `inference`, `disks`, `sandboxes`, `images`).

        `resource_id` is the matching FK value, so callers can drill back into

        the run / pod / disk / etc. that produced the charge.'
    CreateDedicatedRunRequest:
      properties:
        config:
          type: object
          title: Config
          description: Full prime-rl-style TOML, parsed to a dict. Same shape as `prime-rl/examples/*/rl.toml`; the platform
            splits it into trainer / orchestrator / inference subconfigs and bakes each into the corresponding pod's startup
            command.
        imageTag:
          type: string
          title: Imagetag
          description: prime-rl container image tag
          default: main
        name:
          anyOf:
          - type: string
          - type: 'null'
          title: Name
          description: Optional human-readable run name
        teamId:
          anyOf:
          - type: string
          - type: 'null'
          title: Teamid
          description: Owning team (defaults to caller's user)
        wandbApiKey:
          anyOf:
          - type: string
          - type: 'null'
          title: Wandbapikey
          description: W&B key. Materialised into the run's k8s Secret only — never stored on the RFTRun row.
        hfToken:
          anyOf:
          - type: string
          - type: 'null'
          title: Hftoken
          description: HF token for gated/private model downloads.
      type: object
      required:
      - config
      title: CreateDedicatedRunRequest
    RunUsageResponse:
      properties:
        run_id:
          type: string
          title: Run Id
        run_name:
          anyOf:
          - type: string
          - type: 'null'
          title: Run Name
        base_model:
          anyOf:
          - type: string
          - type: 'null'
          title: Base Model
        status:
          anyOf:
          - type: string
          - type: 'null'
          title: Status
        training:
          $ref: '#/components/schemas/RunUsageBreakdown'
        inference:
          $ref: '#/components/schemas/RunUsageBreakdown'
        total_tokens:
          type: integer
          title: Total Tokens
        total_cost_usd:
          type: number
          title: Total Cost Usd
        pricing:
          $ref: '#/components/schemas/RunPricing'
        record_count:
          type: integer
          title: Record Count
          default: 0
      type: object
      required:
      - run_id
      - training
      - inference
      - total_tokens
      - total_cost_usd
      - pricing
      title: RunUsageResponse
      description: Per-run usage totals — drives `prime train usage <run_id>`.
    RunUsageBreakdown:
      properties:
        tokens:
          type: integer
          title: Tokens
          description: Total tokens recorded for this usage type
          default: 0
        input_tokens:
          type: integer
          title: Input Tokens
          description: Total input tokens (only populated for INFERENCE)
          default: 0
        output_tokens:
          type: integer
          title: Output Tokens
          description: Total output tokens (only populated for INFERENCE)
          default: 0
        cost_usd:
          type: number
          title: Cost Usd
          description: Total cost charged for this usage type
          default: 0.0
      type: object
      title: RunUsageBreakdown
      description: Token + cost totals for a single usage type within a run.
    RunPricing:
      properties:
        training_per_mtok:
          anyOf:
          - type: number
          - type: 'null'
          title: Training Per Mtok
        inference_input_per_mtok:
          anyOf:
          - type: number
          - type: 'null'
          title: Inference Input Per Mtok
        inference_output_per_mtok:
          anyOf:
          - type: number
          - type: 'null'
          title: Inference Output Per Mtok
      type: object
      title: RunPricing
      description: Snapshotted per-million-token pricing used to bill the run.
    CreateDedicatedRunResponse:
      properties:
        runId:
          type: string
          title: Runid
        jobId:
          type: string
          title: Jobid
        tokenValue:
          type: string
          title: Tokenvalue
          description: PRIME_API_KEY for this run. Returned once — the platform stores only the token id.
      type: object
      required:
      - runId
      - jobId
      - tokenValue
      title: CreateDedicatedRunResponse
    DeleteRunResponse:
      properties:
        runId:
          type: string
          title: Runid
        deleted:
          type: boolean
          title: Deleted
      type: object
      required:
      - runId
      - deleted
      title: DeleteRunResponse