> ## 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. # Token Prices: Latest > LLM inference prices per model, in USD per 1M tokens. Returns the current value for every record matching your filters — a snapshot of where things stand now, not a series over time. Results are paged. Follow `meta.next` until it is `null`, and send nothing alongside it — the link already carries your filters, dates and ordering. See [Paging](/pricing/concepts#paging). ## Example ```bash theme={null} curl -H "Authorization: Bearer $TOKEN" \ "https://data-api.compute-index.com/v2/llm/token/prices/latest?canonical_model_id=anthropic/claude-opus-4-8" ``` ## OpenAPI ````yaml pricing/openapi.json GET /v2/llm/token/prices/latest openapi: 3.1.0 info: title: GPU Pricing API (v2) description: > Pricing data for GPU compute — rental and purchase prices, settled transactions, published indices, and LLM inference prices. Every data type exposes the same four endpoints over one filter vocabulary, one response envelope and one error format, so learning one is learning all of them. **Start at `/metadata`**, which lists the values a type accepts and the range it holds. <details> <summary><strong>Endpoints</strong> — what the four paths return</summary> | path | returns | | --- | --- | | `/latest` | the current state, one row per thing | | `/as-of` | the state as it stood on a given date | | `/history` | rows over a time window | | `/metadata` | the type's own vocabulary — the valid filter values, and what it covers | Some types have fewer. Settled transactions are events, and "the current state of events" is not a question, so those types serve `/history` and `/metadata` and nothing else. </details> <details> <summary><strong>Responses and paging</strong> — the envelope, and how to get page two</summary> Every response is `{"data": [...], "meta": {...}}`. `data` is the rows; `meta` describes the request they answer — the sort applied, the window resolved, and the paging position. Nulls are always present. A null means the value is not published, which is different from zero and different from absent. `page_size` sets the rows per page (default 100, max 1000). Follow `meta.next` for the following page, and stop when it is null. `meta.next` is an object describing the following page three ways: `url` is an absolute URL ready to call, and `path` plus `query_params` are the same request split up for clients that build their own. Use whichever suits your HTTP layer. Each carries a `page_token`, and the token IS the request — it replays your filters, window and sort exactly — so a following page cannot silently change the question. There is nothing to merge in and nothing to add: do not construct or edit a token, and send no parameters beside it. There is no offset, and no total count. </details> <details> <summary><strong>Filtering, sorting and time</strong> — anchors, sort fields, UTC windows</summary> Filters are per type and listed on each operation. Values are exact matches unless the parameter says otherwise, and where a list is accepted it is comma-separated. **Some types require an anchor** — at least one filter that narrows the query, such as a vendor or a GPU. Asking for everything at once is rejected with a clear error naming the fields that qualify, and `/metadata` lists them. `sort_by` and `sort_order` control ordering, from that type's own set of sortable fields; `meta.sort` reports what was applied. All timestamps are UTC, ISO-8601. Windows are `start`/`end` and are half-open — `start` is included, `end` is not — so consecutive windows tile without double-counting a boundary row. </details> <details> <summary><strong>Access and errors</strong> — tokens, entitlements, rate limits</summary> Send a bearer token: `Authorization: Bearer <token>`. Entitlements are per data type, so a token reaches the types it is granted and 403s on the rest. Rate limits are reported on every response in the `X-RateLimit-*` headers. Errors are RFC 9457 problem details, served as `application/problem+json` with `type`, `title`, `status` and `detail`. </details> contact: name: API Support email: david@compute-index.com version: 2.0.0 servers: - url: https://data-api.compute-index.com security: [] tags: - name: GPU Rental Prices description: >- What it costs to rent GPU compute, as listed by the vendors themselves. One row is one vendor's price for one product. Prices are normalised so they compare across vendors: read `price_usd_per_gpu_hour`. `raw_price` and `raw_price_unit` keep whatever the vendor actually published, for when you need the original. A price that changes produces a new row rather than editing the old one, so `/history` reads as a list of changes and `/as-of` as a single state. Collection queries need an anchor — a vendor, a GPU, or another narrowing filter. `/metadata` lists which fields qualify here. - name: GPU Rental Aggregated Transactions description: >- Settled GPU deals published as blends rather than individually. Every row is a volume-weighted blend of at least two deals, so no single deal can be identified from it. `price` is the only price figure on a row: there is no median, range or spread to ask for. **Do not sum across rows.** Some rows share a deal with the row before them, so adding up `total_gpu_count` or `total_gpu_hours` will double-count. `is_rolling` marks those rows, and `blend_axis` names what was combined to make up the row. There is no `/latest` or `/as-of` — use `/history` with a window. - name: GPU Price Indexes description: >- Published GPU price indices — blended transaction indices at daily and 30-minute resolution, by GPU family or by GPU, per region — and the GPU reference price, the fleet-weighted mean of the per-GPU indices. **One index per request.** Every data endpoint takes a required `id`; `/metadata` lists the ids you hold, and an id you do not hold returns 404. **Two row shapes, told apart by `type`.** A `daily` row's `timestamp` is a calendar date (`2026-09-02`); an `intraday` row's is the UTC instant its interval starts (`2026-09-02T10:00:00+00:00`). `/metadata` gives each index's `row_type`, so you know the shape before you ask. Sort by `timestamp` or `value`. `/metadata` says who computes each index. `p5` and `p95` bracket `value` where an envelope is published, and are null where it is not. - name: GPU Hardware Prices description: >- What it costs to buy GPU hardware outright, from resale, marketplace and reseller listings — the capex counterpart to rental prices. One row is one listing on one day it was seen. A listing that was not seen on a given day has no row for that day, so `/latest` returns each listing's most recent observation rather than a single snapshot. Check `last_observed` in `/metadata` before reading a product's price as current. **Seller identity is not published** — no platform, seller or listing URL, and none of them filterable. What you get instead is the classification a price needs to be read correctly: `market`, `channel`, `price_type`, `source_type`, `confidence` and `counterparty_class`. Check the `includes_*` fields too — a whole-chassis price and a bare-module price are not the same number. - name: LLM Token Prices description: >- Published prices for LLM inference, normalised to USD per 1M tokens so they compare across sources that quote per-token, per-1k and per-1M. **One model has many rows.** A row is identified by the whole combination of source, model, serving provider, price type, region, metric and qualifiers — input and output are priced separately, and the same model appears once per source and once per provider. Filtering to a model and expecting a single row will not work. Implausible prices are flagged rather than removed: `is_price_outlier` marks them and they are included by default, so a response is the data as recorded. Pass `exclude_outliers=true` to drop them. paths: /v2/llm/token/prices/latest: get: tags: - LLM Token Prices summary: Current LLM Token Prices description: >- LLM inference prices per model, in USD per 1M tokens. Returns the current value for every record matching your filters — a snapshot of where things stand now, not a series over time. operationId: latest_llm_token_prices_latest_get parameters: - name: canonical_model_id in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Canonical model id, exact match — `<lab>/<family>` as resolved from the model catalog. Comma-separated list accepted (max 20). Values are listed by this data type's `/metadata`. examples: - anthropic/claude-opus-4-8 title: Canonical Model Id description: >- Canonical model id, exact match — `<lab>/<family>` as resolved from the model catalog. Comma-separated list accepted (max 20). Values are listed by this data type's `/metadata`. - name: base_product_key in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Identity hash: every observation of one priced thing over time, across price changes. Obtain it from a previous response — it is the natural anchor for history drill-down. title: Base Product Key description: >- Identity hash: every observation of one priced thing over time, across price changes. Obtain it from a previous response — it is the natural anchor for history drill-down. - name: product_key in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Snapshot identity hash: one exact observation (identity + time + price). title: Product Key description: >- Snapshot identity hash: one exact observation (identity + time + price). - name: lab in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Model creator, exact match (lower-case). Comma-separated list accepted (max 20). A refinement, not an anchor: pair it with a model id, or list models from `/metadata`. examples: - anthropic title: Lab description: >- Model creator, exact match (lower-case). Comma-separated list accepted (max 20). A refinement, not an anchor: pair it with a model id, or list models from `/metadata`. - name: metric in: query required: false schema: anyOf: - type: string - type: 'null' description: >- What is priced, exact match. Comma-separated list accepted (max 20). Example: `input_text_token,output_text_token` examples: - output_text_token title: Metric description: >- What is priced, exact match. Comma-separated list accepted (max 20). Example: `input_text_token,output_text_token` - name: source in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Aggregator the price was read from, exact match. Comma-separated list accepted. One of `openrouter`, `litellm`, `models_dev`, `artificial_analysis`. title: Source description: >- Aggregator the price was read from, exact match. Comma-separated list accepted. One of `openrouter`, `litellm`, `models_dev`, `artificial_analysis`. - name: price_type in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Kind of price, exact match: `list` (a provider's published catalogue), `marketplace` (a marketplace's own rate), `measured` (observed by a third-party evaluator). Comma-separated list accepted. title: Price Type description: >- Kind of price, exact match: `list` (a provider's published catalogue), `marketplace` (a marketplace's own rate), `measured` (observed by a third-party evaluator). Comma-separated list accepted. - name: provider in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Serving host, exact match (lower-case). Comma-separated list accepted. Null on rows from a marketplace or a measurement rather than a named host. title: Provider description: >- Serving host, exact match (lower-case). Comma-separated list accepted. Null on rows from a marketplace or a measurement rather than a named host. - name: modality in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Model modality, exact match. Comma-separated list accepted. Example: `text,multimodal` title: Modality description: >- Model modality, exact match. Comma-separated list accepted. Example: `text,multimodal` - name: min_price_per_mtok in: query required: false schema: anyOf: - type: number minimum: 0 - type: 'null' description: Only rows priced at >= this many USD/1M tokens (inclusive). title: Min Price Per Mtok description: Only rows priced at >= this many USD/1M tokens (inclusive). - name: max_price_per_mtok in: query required: false schema: anyOf: - type: number minimum: 0 - type: 'null' description: Only rows priced at <= this many USD/1M tokens (inclusive). title: Max Price Per Mtok description: Only rows priced at <= this many USD/1M tokens (inclusive). - name: exclude_outliers in: query required: false schema: type: boolean description: >- Drop rows flagged `is_price_outlier` (>20x the cross-source median for the same model, metric and unit — in practice a source unit error). Default false: the data is served as recorded, with the flag on every row. default: false title: Exclude Outliers description: >- Drop rows flagged `is_price_outlier` (>20x the cross-source median for the same model, metric and unit — in practice a source unit error). Default false: the data is served as recorded, with the flag on every row. - name: sort_by in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Sort field. One of: canonical_model_id, metric, price_per_mtok, source, valid_from. Default: `canonical_model_id`. title: Sort By description: >- Sort field. One of: canonical_model_id, metric, price_per_mtok, source, valid_from. Default: `canonical_model_id`. - name: sort_order in: query required: false schema: anyOf: - type: string - type: 'null' description: 'Sort direction: `asc` or `desc`. Default: `asc`.' title: Sort Order description: 'Sort direction: `asc` or `desc`. Default: `asc`.' - name: page_size in: query required: false schema: type: integer maximum: 1000 minimum: 1 description: Rows per page (max 1000). default: 100 title: Page Size description: Rows per page (max 1000). - name: page_token in: query required: false schema: anyOf: - type: string - type: 'null' description: >- Opaque token token from a previous response's `meta.next`. It carries the whole request — filters, ordering, window and position — so supply it alone; the only parameter that may accompany it is `page_size`. Do not construct or parse these; the format is not part of the contract. title: Page Token description: >- Opaque token token from a previous response's `meta.next`. It carries the whole request — filters, ordering, window and position — so supply it alone; the only parameter that may accompany it is `page_size`. Do not construct or parse these; the format is not part of the contract. responses: '200': description: Successful Response content: application/json: schema: $ref: '#/components/schemas/Page_TokenPriceRow_PageMeta_' '401': description: Unauthorized — `code` is `unauthenticated`. content: application/problem+json: example: status: 401 title: Unauthorized detail: Missing or invalid bearer token. code: unauthenticated request_id: 01JQ2X8N4T6V9WZ0ABCD '403': description: Forbidden — `code` is `insufficient_entitlement`. content: application/problem+json: example: status: 403 title: Forbidden detail: The token does not carry the scopes this endpoint requires. code: insufficient_entitlement request_id: 01JQ2X8N4T6V9WZ0ABCD '404': description: Not Found — `code` is `not_found`. content: application/problem+json: example: status: 404 title: Not Found detail: No such resource, or none you are entitled to see. code: not_found request_id: 01JQ2X8N4T6V9WZ0ABCD '422': description: Unprocessable Content — `code` is `validation_failed`. content: application/problem+json: example: status: 422 title: Unprocessable Content detail: A parameter failed validation. code: validation_failed request_id: 01JQ2X8N4T6V9WZ0ABCD '429': description: Too Many Requests — `code` is `rate_limited`. content: application/problem+json: example: status: 429 title: Too Many Requests detail: Rate limit exceeded; see the X-RateLimit-* headers. code: rate_limited request_id: 01JQ2X8N4T6V9WZ0ABCD security: - HTTPBearer: [] components: schemas: Page_TokenPriceRow_PageMeta_: properties: data: items: $ref: '#/components/schemas/TokenPriceRow' type: array title: Data description: The rows in this page, ordered as `meta.sort` reports. meta: $ref: '#/components/schemas/PageMeta' additionalProperties: false type: object required: - data - meta title: Page[TokenPriceRow, PageMeta] TokenPriceRow: properties: canonical_model_id: anyOf: - type: string - type: 'null' title: Canonical Model Id description: The normalised model identifier, and the anchor to query by. examples: - anthropic/claude-opus-4-8 lab: anyOf: - type: string - type: 'null' title: Lab description: Who trained the model. examples: - anthropic modality: anyOf: - type: string - type: 'null' title: Modality examples: - text is_open_weight: anyOf: - type: boolean - type: 'null' title: Is Open Weight description: Whether the weights are published. context_length: anyOf: - type: integer - type: 'null' title: Context Length description: Context window in tokens. examples: - 200000 tokenizer: anyOf: - type: string - type: 'null' title: Tokenizer description: The tokenizer the price is denominated against. metric: anyOf: - type: string - type: 'null' title: Metric description: What is priced — input, output, cache read. A model has several. examples: - output_text_token price_type: anyOf: - type: string - type: 'null' title: Price Type description: List, promotional, batch, and similar. examples: - list source: anyOf: - type: string - type: 'null' title: Source description: Where the price was collected from. examples: - litellm source_model_id: anyOf: - type: string - type: 'null' title: Source Model Id description: The model id as that source names it. provider: anyOf: - type: string - type: 'null' title: Provider description: Who serves the model, which may not be the lab. region: anyOf: - type: string - type: 'null' title: Region examples: - global qualifiers: anyOf: - additionalProperties: true type: object - type: 'null' title: Qualifiers description: Tier or condition this price applies under; open-ended by source. examples: - {} price_per_mtok: anyOf: - type: number - type: 'null' title: Price Per Mtok description: >- The normalised price: USD per 1M tokens, comparable across sources whatever unit they publish. Zero is a real price — some tiers are free — not a missing value. examples: - 75 raw_price: anyOf: - type: number - type: 'null' title: Raw Price description: The price as the source published it. raw_price_unit: anyOf: - type: string - type: 'null' title: Raw Price Unit description: The unit `raw_price` is in. examples: - USD/token is_price_outlier: anyOf: - type: boolean - type: 'null' title: Is Price Outlier description: >- Flagged, not dropped: a per-token figure published as per-Mtok is 1e6 out, and this marks it. Such rows are included by default, so a response is the data as recorded rather than a filtered view of it; pass `exclude_outliers=true` to exclude them. valid_from: anyOf: - type: string - type: 'null' title: Valid From description: Inclusive start of the interval, UTC instant. valid_to: anyOf: - type: string - type: 'null' title: Valid To description: Exclusive end; null means not yet superseded. extracted_at: anyOf: - type: string - type: 'null' title: Extracted At description: When this snapshot was captured. Not a validity bound. base_product_key: anyOf: - type: string - type: 'null' title: Base Product Key description: Hash identifying this line across its price changes. product_key: anyOf: - type: string - type: 'null' title: Product Key description: Hash identifying this one immutable row. additionalProperties: false type: object required: - canonical_model_id - lab - modality - is_open_weight - context_length - tokenizer - metric - price_type - source - source_model_id - provider - region - qualifiers - price_per_mtok - raw_price - raw_price_unit - is_price_outlier - valid_from - valid_to - extracted_at - base_product_key - product_key title: TokenPriceRow description: >- One priced line of one model, over one validity interval. TALL: identity is the whole tuple of (source, model, provider, price_type, region, metric, qualifiers), so a single model carries many rows — input and output priced separately, per source, per serving provider, per qualifier tier. Filtering to one model and expecting one row is the commonest way to misread this type. PageMeta: properties: page_size: type: integer title: Page Size description: Rows requested per page. examples: - 100 returned: type: integer title: Returned description: Rows in this page, which is `page_size` or fewer. examples: - 100 has_more: type: boolean title: Has More description: Whether a further page exists. next: anyOf: - $ref: '#/components/schemas/NextPage' - type: 'null' description: >- How to fetch the following page, or null at the end of the walk. Use it exactly as given: it locks in every defaulted parameter — the window and the resolved ordering in particular — so later pages cannot drift against earlier ones. There is no `offset`, and nothing may be added to the request it describes. sort: $ref: '#/components/schemas/SortMeta' additionalProperties: false type: object required: - page_size - returned - has_more - next - sort title: PageMeta description: >- What every collection endpoint reports about the page itself. There is no total count. `has_more` tells you whether to fetch another page. NextPage: properties: path: type: string title: Path description: Path to call, without an origin. examples: - /v2/gpu/rental/prices/history query_params: additionalProperties: type: string type: object title: Query Params description: >- The query parameters to send, unencoded. Always exactly the token and the page size — the token carries every filter, the window and the ordering, so nothing else may accompany it. examples: - page_size: '100' page_token: eyJkIjoicHJpY2VzIiwiZSI6Imhpc3RvcnkiLC4uLg.qgLm5w url: type: string title: Url description: The same request as an absolute URL, ready to call. examples: - >- https://data-api.compute-index.com/v2/gpu/rental/prices/history?page_token=eyJkIjoicHJpY2VzIiwiZSI6Imhpc3RvcnksLi4u.qgLm5w&page_size=100 additionalProperties: false type: object required: - path - query_params - url title: NextPage description: >- Where to get the following page, given three ways. All three describe the same request, so use whichever suits your HTTP layer: `url` to call directly, `path` if you already hold a base URL, and `query_params` if you pass parameters as a mapping rather than building a query string. There is nothing to merge in and nothing to add: the page token carries the whole request — every filter, the window and the ordering — so this is the complete next call. SortMeta: properties: by: type: string title: By description: The field the rows are ordered by. examples: - valid_from order: type: string title: Order description: '`asc` or `desc`.' examples: - desc additionalProperties: false type: object required: - by - order title: SortMeta description: |- The ordering actually applied, echoed so a caller who sent no sort parameters can read what they got instead of inferring it from the docs. securitySchemes: HTTPBearer: type: http scheme: bearer ````