> ## Documentation Index
> Fetch the complete documentation index at: https://docs.veri.studio/llms.txt
> Use this file to discover all available pages before exploring further.

# GPU availability by provider/region



## OpenAPI

````yaml /api-reference/openapi.json get /v1/gpu/availability
openapi: 3.1.0
info:
  title: Veri API
  description: >-
    REST API for the Veri RL post-training platform. All requests require a
    Bearer API key (`vk_` prefix).
  license:
    name: ''
  version: 0.1.0
servers:
  - url: https://api.veri.studio
    description: Production
security: []
tags:
  - name: Training jobs
    description: Create, monitor, and manage training jobs.
  - name: Datasets
    description: Upload and connect training datasets.
  - name: Deployments
    description: Serve trained models and run inference.
  - name: Volumes
    description: Persistent file storage mounted into jobs.
  - name: Models
    description: Custom model registry deployments serve from.
  - name: Regions
    description: Discover available launch regions.
  - name: GPU
    description: Live GPU availability by provider and region.
  - name: Code artifacts
    description: Upload custom training script bundles.
  - name: Billing
    description: Credit balance and transaction history.
  - name: API keys
    description: Create and revoke API keys.
  - name: Account
    description: The authenticated caller's identity.
  - name: Settings
    description: Account-level integrations (Weights & Biases).
  - name: Metrics
    description: Prometheus metrics export for your own observability stack.
  - name: Evaluators
    description: 'Evaluators: versioned scoring rules (LLM judge, code, human).'
  - name: Experiments
    description: >-
      Experiments: offline runs of a pinned dataset snapshot through a target,
      scored by pinned evaluators.
  - name: Annotation queues
    description: >-
      Human review: queue traces, threads and experiment items, reserve one at a
      time, score them and feed corrections back into datasets.
  - name: Monitors
    description: >-
      Monitors: evaluators scoring a sampled share of a deployment's live
      traffic.
  - name: Observability
    description: Agent conversations, agents and their traffic.
  - name: Public runs
    description: Unauthenticated reads of runs their owners published to Explore.
paths:
  /v1/gpu/availability:
    get:
      tags:
        - GPU
      summary: GPU availability by provider/region
      operationId: availability
      parameters:
        - name: provider
          in: query
          description: |-
            "aws" | "vast" | "digitalocean". Absent = every probeable
            provider we can reach.
          required: false
          schema:
            type: string
        - name: gpu_type
          in: query
          description: >-
            GPU-type filter (e.g. "H100", "A100-80GB", "MI300X-192GB"). Absent =
            all.
          required: false
          schema:
            type: string
        - name: region
          in: query
          description: |-
            Region filter. AWS: a real region. Vast: "vast-market".
            DigitalOcean: a DO region slug (e.g. "tor1").
          required: false
          schema:
            type: string
      responses:
        '200':
          description: Live GPU availability rows across providers
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/GpuAvailabilityList'
      security:
        - bearerAuth: []
components:
  schemas:
    GpuAvailabilityList:
      type: object
      required:
        - object
        - availability
      properties:
        object:
          type: string
        availability:
          type: array
          items:
            $ref: '#/components/schemas/GpuAvailabilityRow'
    GpuAvailabilityRow:
      type: object
      description: >-
        One availability row. Mirrors the transport `AvailabilityRow` JSON
        (SA-3); the

        control plane passes it through verbatim. Optional fields are
        provider-shaped,

        so `serde(default)` lets a row the transport omitted a field on still
        parse.
      required:
        - provider
        - gpu_type
        - region
        - gpu_count
        - status
      properties:
        provider:
          type: string
        gpu_type:
          type: string
        region:
          type: string
        gpu_count:
          type: integer
          format: int32
          description: >-
            GPUs per launchable unit (Vast smallest chunk / DO droplet cards;
            AWS

            whole-node GPU count).
          minimum: 0
        status:
          type: string
          description: |-
            "AVAILABLE" (live count > 0), "OFFERED" (AWS catalog + spot proxy),
            or "UNAVAILABLE".
        available:
          type:
            - integer
            - 'null'
          format: int64
          description: >-
            Live count (Vast matching offers / DO per-region signal; absent for
            AWS).
          minimum: 0
        spot_score:
          type:
            - integer
            - 'null'
          format: int32
          description: AWS Spot placement score 1–10 proxy (absent for other providers).
        price_per_hr:
          type:
            - number
            - 'null'
          format: double
          description: >-
            Per-GPU USD/hr for this shape. On the wire from the transport this
            is the

            provider's RAW published rate; before returning to the caller the
            endpoint

            marks it up by the per-provider markup so it matches the customer

            price on /v1/cost/rates and never leaks the raw margin.
        detail:
          type:
            - string
            - 'null'
          description: Honest fidelity/units note.
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      bearerFormat: API key
      description: API key with the `vk_` prefix. Create one from the dashboard.

````

This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.