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

# Prices: As Of Date

> Query prices that were active/valid at a point in time. A date-only value means END of that day (UTC) — the daily snapshot convention, matching the txn section's as_of; a full timestamp means that exact instant. Answers reflect best current knowledge of the price book: reprocessed source files and vendor visibility changes can revise historical answers, and prices from a vendor that stops reporting remain valid until closed by the staleness sweep (typically up to ~7 days).

Returns the raw per-offer GPU price records that were active/valid at a point in time — the underlying quotes behind the indices, one row per vendor SKU. Required `as_of_date`: a date-only value means the **end of that day (UTC)** (the daily-snapshot convention, matching the index sections' `as_of`); a full timestamp means that exact instant. Narrow the result with any combination of `vendor`, `gpu_model`, `region`, and `contract_type`, and page through with `page` / `page_size`.

<Note>
  **Premium — historical data.** This endpoint requires the `prices-history` scope, granted separately from `prices` — a token can have current access without history. Contact [david@compute-index.com](mailto:david@compute-index.com) to request access.
</Note>

<Info>
  The interactive playground ("Try it") shows only the **first page** of results — the full result set can run to hundreds of thousands of records. Page through the complete data with `page` / `page_size` (max 1000).
</Info>

<Note>
  Answers reflect the best current knowledge of the price book: reprocessed source files and vendor-visibility changes can revise historical answers, and a quote from a vendor that stops reporting stays valid until closed by the staleness sweep (typically up to \~7 days).
</Note>

## Example

```bash theme={null}
# All H100 prices active on Oct 17 2024
curl -H "Authorization: Bearer $TOKEN" \
  "https://data-api.compute-index.com/v1/prices/as-of?as_of_date=2024-10-17&gpu_model=H100"

# Cheapest AWS on-demand in us-east-1 as of a date
curl -H "Authorization: Bearer $TOKEN" \
  "https://data-api.compute-index.com/v1/prices/as-of?as_of_date=2024-10-17&vendor=aws&region=us-east-1&contract_type=ondemand&sort_by=price_usd_per_gpu_hour&sort_order=asc"
```


## OpenAPI

````yaml pricing/openapi.json GET /v1/prices/as-of
openapi: 3.1.0
info:
  title: GPU Pricing API
  description: |2-

        **GPU Pricing API** provides real-time and historical pricing data
        for GPU compute across 50+ cloud providers and GPU marketplaces.

        ## Features
        - Search current prices across vendors
        - Historical price trends
        - GPU transaction data
        - **Transaction Aggregates** — privacy-safe blends of settled GPU transactions
        - **Informational Price Index** — EWMA-smoothed composite indices by GPU family, region, and contract type
        - **Transaction Price Index** — smoothed indices derived from verified GPU transactions
        - **Custom Index** — EWMA-smoothed H100 sub-model indices by region, contract type, and vendor category
        - Global region coverage (US, EU, APAC)
        - GPU availability tracking
        - Scope-based API key access control

        ## Authentication
        Send your token as a Bearer token in the `Authorization` header on every request:
        ```
        Authorization: Bearer YOUR_TOKEN
        ```
        Example:
        ```bash
        curl -H "Authorization: Bearer $TOKEN" "https://data-api.compute-index.com/v1/vendors"
        ```
        Tokens are issued by Compute Index (david@compute-index.com); long-lived, treat as secrets.

        > The legacy `X-API-Key` header is deprecated — use the `Authorization: Bearer` header above.

        ## Rate Limits
        - **Free tier**: 10 requests/minute, 1,000/day
        - **Pro tier**: 100 requests/minute, 100,000/day
        - **Enterprise**: Custom limits

        Limits are per key, on fixed windows: the per-minute window resets at the top
        of each minute and the daily quota at 00:00 UTC.

        ## Data Coverage
        - 50+ vendors (AWS, Azure, GCP, Oracle, CoreWeave, Lambda, RunPod, Vast.AI, FluidStack, Vultr, etc.)
        - 30M+ historical pricing records
        - Updated multiple times daily

        ## Support
        - Support: david@compute-index.com
        
  contact:
    name: API Support
    email: david@compute-index.com
  license:
    name: Commercial License
    url: https://compute-desk.com/license
  version: 1.0.0
servers:
  - url: https://data-api.compute-index.com
security: []
paths:
  /v1/prices/as-of:
    get:
      tags:
        - Prices
      summary: Get GPU prices as of a specific date
      description: >-
        Query prices that were active/valid at a point in time. A date-only
        value means END of that day (UTC) — the daily snapshot convention,
        matching the txn section's as_of; a full timestamp means that exact
        instant. Answers reflect best current knowledge of the price book:
        reprocessed source files and vendor visibility changes can revise
        historical answers, and prices from a vendor that stops reporting remain
        valid until closed by the staleness sweep (typically up to ~7 days).
      operationId: get_as_of_date_prices_v1_prices_as_of_get
      parameters:
        - name: as_of_date
          in: query
          required: true
          schema:
            type: string
            description: >-
              Target date (ISO format). Date-only = end of that day (UTC); full
              timestamp = exact instant. Returns records where available_at <=
              as_of AND (valid_to IS NULL OR valid_to > as_of). Example:
              2024-10-17 or 2024-10-17T12:00:00
            title: As Of Date
          description: >-
            Target date (ISO format). Date-only = end of that day (UTC); full
            timestamp = exact instant. Returns records where available_at <=
            as_of AND (valid_to IS NULL OR valid_to > as_of). Example:
            2024-10-17 or 2024-10-17T12:00:00
        - name: page
          in: query
          required: false
          schema:
            type: integer
            minimum: 1
            description: Page number (starts at 1)
            default: 1
            title: Page
          description: Page number (starts at 1)
        - name: page_size
          in: query
          required: false
          schema:
            type: integer
            maximum: 1000
            minimum: 1
            description: Items per page (max 1000)
            default: 100
            title: Page Size
          description: Items per page (max 1000)
        - name: vendor
          in: query
          required: false
          schema:
            anyOf:
              - type: string
              - type: 'null'
            description: >-
              Filter by vendor (exact match). Supports comma-separated list (max
              20). Example: aws,gcp,azure
            title: Vendor
          description: >-
            Filter by vendor (exact match). Supports comma-separated list (max
            20). Example: aws,gcp,azure
        - name: gpu_model
          in: query
          required: false
          schema:
            anyOf:
              - type: string
              - type: 'null'
            description: >-
              Filter by GPU model (partial match). Supports comma-separated list
              (max 5). Example: H100,A100,L40S
            title: Gpu Model
          description: >-
            Filter by GPU model (partial match). Supports comma-separated list
            (max 5). Example: H100,A100,L40S
        - name: gpu_canonical_name
          in: query
          required: false
          schema:
            anyOf:
              - type: string
              - type: 'null'
            description: >-
              Filter by canonical GPU name (exact match). Supports
              comma-separated list (max 20). Example: nvidia-h100-sxm-80gb
            title: Gpu Canonical Name
          description: >-
            Filter by canonical GPU name (exact match). Supports comma-separated
            list (max 20). Example: nvidia-h100-sxm-80gb
        - name: region
          in: query
          required: false
          schema:
            anyOf:
              - type: string
              - type: 'null'
            description: >-
              Filter by region (exact match). Supports comma-separated list (max
              20). Example: us-east-1,eu-west-1
            title: Region
          description: >-
            Filter by region (exact match). Supports comma-separated list (max
            20). Example: us-east-1,eu-west-1
        - name: contract_type
          in: query
          required: false
          schema:
            anyOf:
              - type: string
              - type: 'null'
            description: >-
              Filter by contract type. Supports comma-separated list (max 20).
              Example: ondemand,spot,reserved
            title: Contract Type
          description: >-
            Filter by contract type. Supports comma-separated list (max 20).
            Example: ondemand,spot,reserved
        - name: product_key
          in: query
          required: false
          schema:
            anyOf:
              - type: string
              - type: 'null'
            description: Specific product snapshot (includes timestamp and price)
            title: Product Key
          description: Specific product snapshot (includes timestamp and price)
        - name: base_product_key
          in: query
          required: false
          schema:
            anyOf:
              - type: string
              - type: 'null'
            description: Product identity (all historical snapshots of same product)
            title: Base Product Key
          description: Product identity (all historical snapshots of same product)
        - name: sort_by
          in: query
          required: false
          schema:
            type: string
            description: Sort field
            default: price_usd_per_gpu_hour
            title: Sort By
          description: Sort field
        - name: sort_order
          in: query
          required: false
          schema:
            type: string
            description: Sort order (asc or desc)
            default: asc
            title: Sort Order
          description: Sort order (asc or desc)
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/GPUPriceListResponse'
              example:
                data:
                  - availability: available
                    available_at: '2026-07-16T19:01:27.837300+00:00'
                    base_product_key: >-
                      c66b9c50d8d9b2e41e0dfdf307c4799c8c6e3a1c55606586e1f7cbd70b0d1011
                    contract_details: 'max_duration: 17.8 days'
                    contract_term: null
                    contract_type: ondemand
                    created_by: ingest_incremental.py
                    extracted_at: '2026-07-16T19:01:27.837300+00:00'
                    form_factor: SXM
                    gpu_canonical_name: nvidia-h100-sxm5-80gb
                    gpu_count: 2
                    gpu_manufacturer: NVIDIA
                    gpu_memory_gb: 80
                    gpu_model: H100
                    id: 53099890-d175-44f8-80b1-7876c38d095e
                    instance_type: H100 SXM
                    instance_variant: machine:59377, host:456754
                    interconnect: NVLink
                    job_execution_id: null
                    last_seen_at: '2026-07-16T19:01:27.837300+00:00'
                    metadata:
                      availability_unit: gpu
                      available_units: 16
                      bundle_id: 1914300712
                      cpu_ghz: 3.1
                      cpu_name: AMD EPYC 9554 64-Core Processor
                      driver_version: 550.163.01
                      flops_per_dollar: 30.860906598334406
                      gpu_ids:
                        - 182525
                        - 182526
                      inet_down_mbps: 623.6
                      inet_up_mbps: 725.1
                      listing_id: 42557071
                      machine_id: 59377
                      max_duration_display: 17.8 days
                      max_duration_seconds: 1540738.4754431248
                      nvlink_bw_gbps: 478.1160888671875
                      pcie_bw_gbps: 36.8
                      pcie_gen: 5
                      performance_score: 187.5258564098057
                      port_count: 199
                      provider_id: 456754
                      public_ip: 66.189.182.26
                      reliability_score: 0.997798
                      rented_units: 36
                      storage_bw_mbps: 1468.1168334878566
                      storage_gb: 24.5
                      storage_type: M.2 NVMe 2-Bay RAID Kit
                      total_units: 52
                      verification_status: verified
                    os: Linux
                    price_hash: >-
                      fa0cbdbc2c9aac004d841da4fda23f6c421feb27ab3b757b8539df1773cb186f
                    price_usd_per_gpu_hour: 1.734
                    product_key: >-
                      75999c133d285a6f4d029a228a77b386e14e27fe985ad6ccd561e9838f6aa000
                    ram_gb: 252
                    raw_price: 3.469
                    raw_price_unit: USD/hour (total)
                    region: us-east-1
                    source: api
                    source_file: vastai_pricing_20260716_190124_33587.json
                    updated_by: ingest_incremental.py:phase5
                    valid_to: '2026-07-17T00:02:41.224361+00:00'
                    vcpu: 256
                    vendor: vastai
                    vendor_gpu_name: H100 SXM
                    vendor_region: Washington, US
                    vendor_sku: 42557071:h100 sxm:2gpu:ondemand
                  - availability: available
                    available_at: '2026-07-16T19:01:27.851627+00:00'
                    base_product_key: >-
                      cd56cec7b3d1d8e24e69eecf06bc1744463da1a59141b9e6bcf52b0f169a920e
                    contract_details: 'max_duration: 17.8 days'
                    contract_term: null
                    contract_type: ondemand
                    created_by: ingest_incremental.py
                    extracted_at: '2026-07-16T19:01:27.851627+00:00'
                    form_factor: SXM
                    gpu_canonical_name: nvidia-h100-sxm5-80gb
                    gpu_count: 1
                    gpu_manufacturer: NVIDIA
                    gpu_memory_gb: 80
                    gpu_model: H100
                    id: a4dad384-2006-4b32-bb5d-4c02bdbbce9a
                    instance_type: H100 SXM
                    instance_variant: machine:139170, host:456754
                    interconnect: NVLink
                    job_execution_id: null
                    last_seen_at: '2026-07-16T19:01:27.851627+00:00'
                    metadata:
                      availability_unit: gpu
                      available_units: 16
                      bundle_id: 1781148153
                      cpu_ghz: 3.800241
                      cpu_name: AMD EPYC 9354 32-Core Processor
                      driver_version: 550.163.01
                      flops_per_dollar: 30.841149295774652
                      gpu_ids:
                        - 202487
                      inet_down_mbps: 691.6
                      inet_up_mbps: 764.1
                      listing_id: 44760325
                      machine_id: 139170
                      max_duration_display: 17.8 days
                      max_duration_seconds: 1540737.4754431248
                      nvlink_bw_gbps: 478.1160888671875
                      pcie_bw_gbps: 36.8
                      pcie_gen: 5
                      performance_score: 157.2484748289074
                      port_count: 99
                      provider_id: 456754
                      public_ip: 66.189.182.26
                      reliability_score: 0.980523
                      rented_units: 36
                      storage_bw_mbps: 885
                      storage_gb: 32.25
                      storage_type: Micron_7450_MTFDKBA480TFR
                      total_units: 52
                      verification_status: verified
                    os: Linux
                    price_hash: >-
                      0a0b800d35565bac1b1f3a2206b6323c431b89142291fcbc70ba76ae35641c9c
                    price_usd_per_gpu_hour: 1.736
                    product_key: >-
                      902ec5f993acc691453e95787c9f6c8ac49758263330c68a56b158526f28585a
                    ram_gb: 126
                    raw_price: 1.736
                    raw_price_unit: USD/hour (total)
                    region: us-east-1
                    source: api
                    source_file: vastai_pricing_20260716_190124_33587.json
                    updated_by: ingest_incremental.py:phase5
                    valid_to: '2026-07-17T00:02:41.224361+00:00'
                    vcpu: 128
                    vendor: vastai
                    vendor_gpu_name: H100 SXM
                    vendor_region: Washington, US
                    vendor_sku: 44760325:h100 sxm:1gpu:ondemand
                filters:
                  as_of_date: '2026-07-16'
                  base_product_key: null
                  contract_type: ondemand
                  gpu_canonical_name: null
                  gpu_model: H100
                  product_key: null
                  region: null
                  sort_by: price_usd_per_gpu_hour
                  sort_order: asc
                  vendor: null
                pagination:
                  has_next: true
                  has_previous: false
                  page: 1
                  page_size: 100
                  total_count: 1639
                  total_pages: 17
        '400':
          description: Bad Request — malformed date
          content:
            application/json:
              schema:
                title: ErrorResponse
                type: object
                properties:
                  detail:
                    type: string
                required:
                  - detail
              examples:
                invalid_date:
                  summary: Malformed as_of_date
                  value:
                    detail: 'Invalid isoformat string: ''not-a-date'''
        '401':
          description: Unauthorized — missing or invalid credentials
          content:
            application/json:
              schema:
                title: ErrorResponse
                type: object
                properties:
                  detail:
                    type: string
                required:
                  - detail
              examples:
                missing_credentials:
                  summary: No credentials supplied
                  value:
                    detail: >-
                      Authentication required. Provide an X-API-Key header or an
                      Authorization: Bearer token.
                invalid_token:
                  summary: Invalid or expired token
                  value:
                    detail: Invalid or expired token
        '403':
          description: Forbidden — key lacks the required scope
          content:
            application/json:
              schema:
                title: ErrorResponse
                type: object
                properties:
                  detail:
                    type: string
                required:
                  - detail
              examples:
                missing_scope:
                  summary: Key lacks the required scope
                  value:
                    detail: >-
                      Access denied. Your API key does not have permission to
                      access this resource.
        '422':
          description: Validation Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/HTTPValidationError'
              example:
                detail:
                  - type: missing
                    loc:
                      - query
                      - as_of_date
                    msg: Field required
                    input: null
        '429':
          description: >-
            Too Many Requests — limit and reset are returned in X-RateLimit-*
            headers; the limit reflects your key's tier
          content:
            application/json:
              schema:
                title: ErrorResponse
                type: object
                properties:
                  detail:
                    type: string
                required:
                  - detail
              examples:
                minute_limit:
                  summary: Per-minute limit exceeded (figure reflects your key's tier)
                  value:
                    detail: 'Rate limit exceeded: 100 requests per minute'
      security:
        - HTTPBearer: []
components:
  schemas:
    GPUPriceListResponse:
      properties:
        data:
          items:
            $ref: '#/components/schemas/GPUPriceResponse'
          type: array
          title: Data
        pagination:
          additionalProperties: true
          type: object
          title: Pagination
        filters:
          additionalProperties: true
          type: object
          title: Filters
      type: object
      required:
        - data
        - pagination
        - filters
      title: GPUPriceListResponse
      description: Response schema for list of GPU prices with pagination
    HTTPValidationError:
      properties:
        detail:
          items:
            $ref: '#/components/schemas/ValidationError'
          type: array
          title: Detail
      type: object
      title: HTTPValidationError
    GPUPriceResponse:
      properties:
        id:
          type: string
          title: Id
        product_key:
          type: string
          title: Product Key
        price_hash:
          type: string
          title: Price Hash
        base_product_key:
          type: string
          title: Base Product Key
        extracted_at:
          type: string
          format: date-time
          title: Extracted At
        available_at:
          type: string
          format: date-time
          title: Available At
        valid_to:
          anyOf:
            - type: string
              format: date-time
            - type: 'null'
          title: Valid To
        last_seen_at:
          type: string
          format: date-time
          title: Last Seen At
        vendor:
          type: string
          title: Vendor
        vendor_sku:
          type: string
          title: Vendor Sku
        vendor_region:
          type: string
          title: Vendor Region
        region:
          type: string
          title: Region
        contract_type:
          type: string
          title: Contract Type
        vendor_gpu_name:
          type: string
          title: Vendor Gpu Name
        gpu_canonical_name:
          type: string
          title: Gpu Canonical Name
        gpu_manufacturer:
          type: string
          title: Gpu Manufacturer
        gpu_model:
          type: string
          title: Gpu Model
        gpu_count:
          anyOf:
            - type: number
            - type: 'null'
          title: Gpu Count
        gpu_memory_gb:
          anyOf:
            - type: number
            - type: 'null'
          title: Gpu Memory Gb
        form_factor:
          anyOf:
            - type: string
            - type: 'null'
          title: Form Factor
        interconnect:
          anyOf:
            - type: string
            - type: 'null'
          title: Interconnect
        instance_type:
          anyOf:
            - type: string
            - type: 'null'
          title: Instance Type
        instance_variant:
          anyOf:
            - type: string
            - type: 'null'
          title: Instance Variant
        vcpu:
          anyOf:
            - type: number
            - type: 'null'
          title: Vcpu
        ram_gb:
          anyOf:
            - type: number
            - type: 'null'
          title: Ram Gb
        os:
          anyOf:
            - type: string
            - type: 'null'
          title: Os
        price_usd_per_gpu_hour:
          type: number
          title: Price Usd Per Gpu Hour
        raw_price:
          type: number
          title: Raw Price
        raw_price_unit:
          type: string
          title: Raw Price Unit
        contract_term:
          anyOf:
            - type: string
            - type: 'null'
          title: Contract Term
        contract_details:
          anyOf:
            - type: string
            - type: 'null'
          title: Contract Details
        availability:
          type: string
          title: Availability
        source:
          type: string
          title: Source
        metadata:
          anyOf:
            - additionalProperties: true
              type: object
            - type: 'null'
          title: Metadata
        created_by:
          anyOf:
            - type: string
            - type: 'null'
          title: Created By
        updated_by:
          anyOf:
            - type: string
            - type: 'null'
          title: Updated By
        job_execution_id:
          anyOf:
            - type: string
            - type: 'null'
          title: Job Execution Id
        source_file:
          anyOf:
            - type: string
            - type: 'null'
          title: Source File
      type: object
      required:
        - id
        - product_key
        - price_hash
        - base_product_key
        - extracted_at
        - available_at
        - last_seen_at
        - vendor
        - vendor_sku
        - vendor_region
        - region
        - contract_type
        - vendor_gpu_name
        - gpu_canonical_name
        - gpu_manufacturer
        - gpu_model
        - price_usd_per_gpu_hour
        - raw_price
        - raw_price_unit
        - availability
        - source
      title: GPUPriceResponse
      description: Response schema for a single GPU price record
    ValidationError:
      properties:
        loc:
          items:
            anyOf:
              - type: string
              - type: integer
          type: array
          title: Location
        msg:
          type: string
          title: Message
        type:
          type: string
          title: Error Type
        input:
          title: Input
        ctx:
          type: object
          title: Context
      type: object
      required:
        - loc
        - msg
        - type
      title: ValidationError
  securitySchemes:
    HTTPBearer:
      type: http
      scheme: bearer

````