> ## Documentation Index > Fetch the complete documentation index at: https://veryfront.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # List models > List models. ## OpenAPI ````yaml /cloud/rest/openapi.json get /ai/models openapi: 3.1.0 info: title: Veryfront REST API reference version: 1.0.0 description: |- ## Explore the API | Domain | Use it to | | --- | --- | | [Projects](#tag/projects) | Manage projects and files. | | [Agents](#tag/agents) | Manage agent primitives. | | [Orchestration](#tag/execution) | Manage conversations and runs. | | [Evaluations](#tag/evaluations) | Manage evaluations and results. | | [AI Gateway](#tag/ai-gateway) | Access models and inference. | | [MCP](#tag/mcp) | Discover tools and authorization. | | [Integrations](#tag/integrations) | Manage connections and integrations. | | [Knowledge](#tag/knowledge) | Manage documents and search. | | [Deployments](#tag/deployments) | Manage releases and deployments. | | [Runtime](#tag/runtime) | Manage environments and infrastructure. | | [Observability](#tag/observability) | Monitor activity and performance. | | [Identity & Access](#tag/identity-access) | Manage identity and access. | | [Billing & Usage](#tag/billing-usage) | Manage billing and usage. | | [Other](#tag/platform) | Inspect health and API metadata. | [OpenAPI specification (JSON)](/openapi.json) contact: name: Veryfront API Support url: https://veryfront.com/docs email: support@veryfront.com x-api-id: 5f1f0b8a-6b0e-4c1a-9b35-0f3a1e6d84c7 x-audience: external-public servers: - url: https://api.veryfront.com security: - bearerAuth: [] - apiKeyAuth: [] tags: - name: projects description: Manage projects and files. - name: agents description: Manage agent primitives. - name: execution description: Manage conversations and runs. - name: evaluations description: Manage evaluations and results. - name: ai-gateway description: Access models and inference. - name: mcp description: Discover tools and authorization. - name: integrations description: Manage connections and integrations. - name: knowledge description: Manage documents and search. - name: deployments description: Manage releases and deployments. - name: runtime description: Manage environments and infrastructure. - name: observability description: Monitor activity and performance. - name: identity-access description: Manage identity and access. - name: billing-usage description: Manage billing and usage. - name: platform description: Inspect health and API metadata. - name: projects/Projects - name: projects/Projects/Favorites - name: projects/Projects/Account Projects - name: projects/Projects/Projects - name: projects/Projects/Project Labels - name: projects/Projects/Project templates - name: projects/Files - name: projects/Files/Files - name: projects/Files/File Chunks - name: projects/Files/Branch File Versions - name: projects/Files/Branch Files - name: projects/Files/Environment Files - name: projects/Files/File Versions - name: projects/Files/Branch File Drafts - name: projects/Files/Release Files - name: projects/Files/Sandbox Files - name: projects/Files/File Search - name: projects/Branches - name: projects/Branches/Branches - name: projects/Branches/Branch Policies - name: projects/Branches/Branch Approvals - name: projects/Branches/Branch Status Checks - name: projects/Branches/Project Commits - name: projects/Branches/Branch Merges - name: projects/Uploads - name: projects/Uploads/Upload Folders - name: projects/Uploads/Upload Grants - name: projects/Uploads/Project Uploads - name: projects/Uploads/Upload Visibility - name: agents/Agents - name: agents/Agents/Agents - name: agents/Agents/Agent Labels - name: agents/Agents/Agent templates - name: agents/Agents/Agent Tool References - name: agents/Agents/Agent Avatars - name: agents/Agents/Agent Capabilities - name: agents/Agents/Agent Installation - name: agents/Prompts - name: agents/Prompts/Project prompts - name: agents/Tools - name: agents/Tools/Project tools - name: agents/Skills - name: agents/Skills/Project skills - name: agents/Resources - name: agents/Resources/Project resources - name: agents/Tasks - name: agents/Tasks/Tasks - name: agents/Workflows - name: agents/Workflows/Workflows - name: execution/Conversations - name: execution/Conversations/Conversations - name: execution/Conversations/Conversation messages - name: execution/Conversations/Conversation participants - name: execution/Conversations/Conversation read state - name: execution/Triggers - name: execution/Triggers/Project schedules - name: execution/Triggers/Webhooks - name: execution/Triggers/Webhook events - name: execution/Runs - name: execution/Runs/Run Input Requests - name: execution/Runs/Runs - name: execution/Runs/Runtime - name: execution/Runs/Child Runs - name: execution/Runs/Conversation Runs - name: execution/Runs/Project Runs - name: execution/Runs/Webhook Runs - name: execution/Runs/Run Events - name: execution/Runs/Run Actions - name: execution/Runs/Run Artifacts - name: execution/Executors - name: execution/Executors/Connected agents - name: execution/Executors/Agent workers - name: evaluations/Evaluations - name: evaluations/Evaluations/Evaluations - name: evaluations/Evaluations/Evaluation Runs - name: ai-gateway/Models - name: ai-gateway/Models/Model discovery - name: ai-gateway/Models/Inference Models - name: ai-gateway/Models/Project inference policy - name: ai-gateway/Models/Account inference policy - name: ai-gateway/Inference - name: ai-gateway/Inference/OpenAI Chat Completions - name: ai-gateway/Inference/Embeddings - name: ai-gateway/Inference/OpenAI Responses - name: ai-gateway/Inference/Proxy metadata - name: ai-gateway/Inference/Anthropic Messages - name: ai-gateway/Inference/Provider gateway - name: ai-gateway/Inference/Gemini inference - name: ai-gateway/Inference/Image generation - name: mcp/Tools - name: mcp/Tools/MCP discovery - name: mcp/Authorization - name: mcp/Authorization/Protected resource discovery - name: integrations/Integrations - name: integrations/Integrations/Integration tools - name: integrations/Integrations/Integration catalog - name: integrations/Integrations/Project integrations - name: integrations/Connections - name: integrations/Connections/Connections - name: integrations/Channels - name: integrations/Channels/Channels - name: integrations/Channels/Setup lookups - name: integrations/External storage - name: integrations/External storage/External Files - name: knowledge/Documents - name: knowledge/Documents/Project knowledge - name: knowledge/Documents/Retrieval documents - name: knowledge/Search - name: knowledge/Search/Search - name: knowledge/Embeddings - name: knowledge/Embeddings/Project embeddings - name: deployments/Releases - name: deployments/Releases/Releases - name: deployments/Releases/Release Versions - name: deployments/Releases/Release Source Manifests - name: deployments/Deployments - name: deployments/Deployments/Project deployments - name: deployments/Artifacts - name: deployments/Artifacts/Dependencies - name: deployments/Artifacts/Release Asset Manifests - name: deployments/Artifacts/Style artifacts - name: runtime/Environments - name: runtime/Environments/Environments - name: runtime/Environments/Environment Domains - name: runtime/Environments/Environment variables - name: runtime/Environments/Project Domains - name: runtime/Environments/Environment Variable Keys - name: runtime/Servers - name: runtime/Servers/Environment Servers - name: runtime/Servers/Project servers - name: runtime/Sandboxes - name: runtime/Sandboxes/Sandbox Lifecycle - name: runtime/Sandboxes/Sandbox Commands - name: runtime/Sandboxes/Sandbox Environment - name: runtime/Sandboxes/Sandbox Health - name: runtime/Sandboxes/Sandbox Terminal - name: runtime/Cache - name: runtime/Cache/Project Cache - name: runtime/Cache/User Cache - name: observability/Logs - name: observability/Logs/Log entries - name: observability/Traces - name: observability/Traces/Trace spans - name: observability/Traces/Project traces - name: observability/Traces/Trace Search - name: observability/Metrics - name: observability/Metrics/Metric series - name: observability/Alerts - name: observability/Alerts/Project alerts - name: observability/Error types - name: observability/Error types/Error types - name: identity-access/Users - name: identity-access/Users/Account users - name: identity-access/Users/Project users - name: identity-access/Users/Current user - name: identity-access/Accounts - name: identity-access/Accounts/Accounts - name: identity-access/Invitations - name: identity-access/Invitations/Invitations - name: identity-access/Permissions - name: identity-access/Permissions/Feature gates - name: identity-access/Permissions/Project Ownership - name: identity-access/API keys - name: identity-access/API keys/API keys - name: identity-access/Authentication - name: identity-access/Authentication/Sign-in - name: identity-access/Authentication/Tokens - name: identity-access/Authorization - name: identity-access/Authorization/Authorized apps - name: identity-access/Authorization/General OAuth - name: identity-access/Authorization/Consent - name: billing-usage/Subscriptions - name: billing-usage/Subscriptions/Subscriptions - name: billing-usage/Subscriptions/Seats - name: billing-usage/Billing - name: billing-usage/Billing/Invoices - name: billing-usage/Billing/Payments - name: billing-usage/Billing/Payment Methods - name: billing-usage/Billing/Billing Details - name: billing-usage/Billing/Account billing - name: billing-usage/Credits - name: billing-usage/Credits/Credits - name: billing-usage/Budgets - name: billing-usage/Budgets/Budgets - name: billing-usage/Usage - name: billing-usage/Usage/API key usage - name: billing-usage/Usage/Account usage - name: billing-usage/Usage/Project usage - name: billing-usage/Analytics - name: billing-usage/Analytics/Conversation Analytics - name: billing-usage/Analytics/Run Analytics - name: platform/Other - name: platform/Other/API metadata paths: /ai/models: get: tags: - ai-gateway/Models/Model discovery summary: List models description: List models. operationId: listModels parameters: - schema: type: string minLength: 1 maxLength: 255 description: Project UUID or slug used to attribute billing finalization. required: false name: x-veryfront-project-reference in: header - schema: type: string minLength: 1 maxLength: 255 description: Project ID used to attribute billing finalization. required: false name: x-veryfront-project-id in: header - schema: type: string minLength: 1 maxLength: 255 description: Project slug used to attribute billing finalization. required: false name: x-veryfront-project-slug in: header responses: '200': description: >- Available AI models and providers. A project-scoped credential, or a selector header naming a project the caller can see, narrows the list to what that project's inference policy permits. content: application/json: schema: type: object properties: models: type: array items: type: object properties: id: type: string description: Model identifier. example: claude-sonnet-4-6 modelId: type: string description: Unified model ID (provider/model). example: anthropic/claude-sonnet-4-6 provider: type: string description: Provider name. example: anthropic providerLabel: type: string description: Public provider display name. example: Anthropic providerLogoKey: type: string description: Public provider logo key. example: anthropic surface: type: string description: >- Gateway protocol for the model’s provider under `/ai/gateway/{provider}`. Known values are `openai`, `anthropic`, and `google`; clients must accept unknown values. Omitted if no gateway protocol is configured, in which case the gateway accepts no paths for the provider. example: anthropic operations: type: array items: type: string description: >- Operations of `surface` this model is served on, each at its canonical path (for Google, the `models/:` form). Known values: `chat-completions`, `responses`, `embeddings` on `openai`; `messages` on `anthropic`; `generate-content`, `stream-generate-content`, `embed-content`, `batch-embed-contents`, `predict`, `openai-responses`, `openai-chat-completions`, `openai-embeddings` on `google`. Not a closed enum, tolerate values you have not seen. Present exactly when `surface` is. example: - messages name: type: string description: Display name. example: Claude Sonnet description: type: string description: Short description. example: Most efficient for everyday tasks aliases: type: array items: type: string description: >- Accepted aliases for backward-compatible model selection. example: - sonnet - anthropic/claude-sonnet-4-6 capabilities: type: object properties: webResearch: type: boolean description: >- Whether Studio default chat can enable live web research. streaming: type: boolean description: Whether streaming responses are supported. tools: type: boolean description: Whether tool calls are supported. vision: type: boolean description: Whether image inputs are supported. thinking: type: boolean description: >- Whether thinking/reasoning controls are supported. supported_parameters: type: array items: type: string description: >- Supported request parameters, using OpenRouter names. Parameters not listed are removed or rejected. Parameters are listed only when support has been verified. example: - max_tokens - temperature - top_p - tools - tool_choice reasoning: type: boolean description: Whether the model reasons before answering. tool_call: type: boolean description: Whether the model can call tools. temperature: type: boolean description: >- Whether a caller-set temperature reaches the model. open_weights: type: boolean description: Whether the model weights are publicly released. attachment: type: boolean description: >- Whether file input is supported. Omitted if support has not been verified. structured_output: type: boolean description: >- Whether schema-constrained output is supported. Omitted if support has not been verified. modalities: type: object properties: input: type: array items: type: string description: Input modalities. example: - text - image required: - input description: >- Supported input modalities. Output modalities are omitted because they have not been verified. limit: type: object properties: context: type: integer exclusiveMinimum: 0 description: >- Context window in tokens: the provider's documented limit for the deployment Veryfront serves. Input plus requested output must fit within it. Absent while unevidenced. example: 1000000 output: type: integer exclusiveMinimum: 0 description: >- Maximum output tokens. Absent while unevidenced. example: 128000 description: >- Known token ceilings. Each is absent while unevidenced; `limit` is absent while neither is known. reasoning_mode: type: string description: >- Supported reasoning control. Known values are `budget` and `adaptive`. Omitted if the model does not reason or support has not been verified. Clients must accept unknown values. example: budget transport: type: string description: >- OpenAI wire API the client must use. Known values: `chat-completions`, `responses`. Absent when unconstrained. Not a closed enum. example: chat-completions reasoning_budget_tokens: type: integer exclusiveMinimum: 0 description: >- Default thinking budget in tokens when the caller does not specify one. Omitted if no default is configured or the model does not accept a budget, including models without reasoning and models with adaptive thinking. example: 2048 chat_completions_reasoning_with_function_tools: type: boolean description: >- Whether `chat-completions` supports reasoning and function tools in the same request. If false, choose one. Omitted if support has not been verified. example: false chat_completions_consecutive_system_messages: type: boolean description: >- Whether `chat-completions` accepts consecutive system messages. If true, clients can send system instructions separately. Omitted if support has not been verified. example: true required: - webResearch - streaming - tools - vision - supported_parameters - reasoning - tool_call - temperature - open_weights - modalities description: >- Model capability flags. The camelCase keys are the original Veryfront flags; the snake_case keys are the de facto names clients already parse (OpenRouter, models.dev) and are derived from the capability records. temperature: type: - object - 'null' properties: min: type: number minimum: 0 description: Minimum supported temperature. max: type: number minimum: 0 description: Maximum supported temperature. required: - min - max description: >- Per-model temperature limits; null when the runtime fixes or omits temperature. supportedProviderTools: type: array items: type: string description: >- Provider-executed agent tools supported on the deployment a request for this model is served from. example: - web_search - web_fetch pricing: type: - object - 'null' properties: inputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Input-token price, USD per 1M tokens. Use `inputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md outputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Output-token price, USD per 1M tokens. Use `outputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreationInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Default-retention cache-write input-token price, USD per 1M tokens. Use `cacheCreationInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreation1hInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. One-hour cache-write input-token price, USD per 1M tokens. Use `cacheCreation1hInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Cache-read input-token price, USD per 1M tokens. Use `cacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio input-token price, USD per 1M tokens. Use `audioInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioOutputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio output-token price, USD per 1M tokens. Use `audioOutputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioCacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio cache-read token price, USD per 1M tokens. Use `audioCacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md inputEurPer1M: type: number minimum: 0 description: Input-token price. EUR per 1M tokens; see `fx`. outputEurPer1M: type: number minimum: 0 description: Output-token price. EUR per 1M tokens; see `fx`. cacheCreationInputEurPer1M: type: number minimum: 0 description: >- Default-retention cache-write input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. cacheCreation1hInputEurPer1M: type: number minimum: 0 description: >- One-hour cache-write input-token price, when the provider bills it apart from the default cache write. EUR per 1M tokens; see `fx`. cacheReadInputEurPer1M: type: number minimum: 0 description: >- Cache-read input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioInputEurPer1M: type: number minimum: 0 description: >- Audio input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioOutputEurPer1M: type: number minimum: 0 description: >- Audio output-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioCacheReadInputEurPer1M: type: number minimum: 0 description: >- Audio cache-read token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. required: - inputUsdPer1M - outputUsdPer1M - inputEurPer1M - outputEurPer1M description: >- Public Veryfront customer pricing after markup, used for client-side customer cost estimates. Null for externally managed local inference with no evidenced tariff. providerPricing: type: - object - 'null' properties: inputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Input-token price, USD per 1M tokens. Use `inputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md outputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Output-token price, USD per 1M tokens. Use `outputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreationInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Default-retention cache-write input-token price, USD per 1M tokens. Use `cacheCreationInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreation1hInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. One-hour cache-write input-token price, USD per 1M tokens. Use `cacheCreation1hInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Cache-read input-token price, USD per 1M tokens. Use `cacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio input-token price, USD per 1M tokens. Use `audioInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioOutputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio output-token price, USD per 1M tokens. Use `audioOutputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioCacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio cache-read token price, USD per 1M tokens. Use `audioCacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md inputEurPer1M: type: number minimum: 0 description: Input-token price. EUR per 1M tokens; see `fx`. outputEurPer1M: type: number minimum: 0 description: Output-token price. EUR per 1M tokens; see `fx`. cacheCreationInputEurPer1M: type: number minimum: 0 description: >- Default-retention cache-write input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. cacheCreation1hInputEurPer1M: type: number minimum: 0 description: >- One-hour cache-write input-token price, when the provider bills it apart from the default cache write. EUR per 1M tokens; see `fx`. cacheReadInputEurPer1M: type: number minimum: 0 description: >- Cache-read input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioInputEurPer1M: type: number minimum: 0 description: >- Audio input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioOutputEurPer1M: type: number minimum: 0 description: >- Audio output-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioCacheReadInputEurPer1M: type: number minimum: 0 description: >- Audio cache-read token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. tiers: type: array items: type: object properties: inputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Input-token price, USD per 1M tokens. Use `inputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md outputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Output-token price, USD per 1M tokens. Use `outputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreationInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Default-retention cache-write input-token price, USD per 1M tokens. Use `cacheCreationInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreation1hInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. One-hour cache-write input-token price, USD per 1M tokens. Use `cacheCreation1hInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Cache-read input-token price, USD per 1M tokens. Use `cacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio input-token price, USD per 1M tokens. Use `audioInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioOutputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio output-token price, USD per 1M tokens. Use `audioOutputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioCacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio cache-read token price, USD per 1M tokens. Use `audioCacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md inputEurPer1M: type: number minimum: 0 description: >- Input-token price. EUR per 1M tokens; see `fx`. outputEurPer1M: type: number minimum: 0 description: >- Output-token price. EUR per 1M tokens; see `fx`. cacheCreationInputEurPer1M: type: number minimum: 0 description: >- Default-retention cache-write input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. cacheCreation1hInputEurPer1M: type: number minimum: 0 description: >- One-hour cache-write input-token price, when the provider bills it apart from the default cache write. EUR per 1M tokens; see `fx`. cacheReadInputEurPer1M: type: number minimum: 0 description: >- Cache-read input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioInputEurPer1M: type: number minimum: 0 description: >- Audio input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioOutputEurPer1M: type: number minimum: 0 description: >- Audio output-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioCacheReadInputEurPer1M: type: number minimum: 0 description: >- Audio cache-read token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. id: type: string description: Stable provider pricing tier identifier example: long-context label: type: string description: Human-readable pricing tier label example: Long context appliesTo: type: string description: >- Condition where this provider pricing tier applies. example: Prompts over 200k tokens required: - inputUsdPer1M - outputUsdPer1M - inputEurPer1M - outputEurPer1M - id - label - appliesTo description: >- Raw provider pricing tier that is not the default catalog billing rate. description: >- Additional raw provider pricing tiers not currently represented by the default public price. required: - inputUsdPer1M - outputUsdPer1M - inputEurPer1M - outputEurPer1M description: >- What the provider charges Veryfront (our upstream cost), before markup. For a host that invoices in EUR the EUR prices are that host's published EUR list, unconverted and unrounded. pricingSources: type: - object - 'null' properties: directProvider: type: string format: uri description: Official provider pricing source URL. azure: type: string format: uri description: Official Azure-hosted model pricing source URL. required: - directProvider description: Official pricing source links for this model. deployments: type: array items: type: object properties: tier: type: string description: >- Residency tier of this deployment. Known values: `qualified-sovereign`, `eu-owned`, `eu-resident`, `standard`. example: eu-resident jurisdiction: type: string description: >- Legal domicile (ISO 3166-1 alpha-2) of the company operating the inference, or `unknown` where unevidenced. This is the operator's home country, NOT where inference is processed: an EU-resident deployment run by a US company reads `US`. Read `processing` for the location. example: US region: type: string description: >- Where inference is PROCESSED, not where the cloud resource was created. `eu` is an EU boundary (the EU Data Zone or an EU multi-region, depending on the host), `europe-west1` / `europe-west4` are single EU member-state regions, `global` means the route is worldwide and no region may be claimed, and `unknown` means unevidenced. example: europe-west1 processing: type: string description: >- Where inference is processed, normalised from `region` and `tier`. Known values: `eu` (inside a recorded EU boundary), `us` (in the United States), `worldwide` (in any region the operator runs), `unknown` (no processing region is recorded for this route). An open string: tolerate values you have not seen. example: eu processing_label: type: string description: >- The processing location in words, for display, e.g. `EU (EEA + Switzerland data zone)`, `EU (europe-west1, Belgium)`, `United States`, `Worldwide (any region)`, `Unknown (processing region not recorded)`. Display text only: branch on `processing`. example: EU (europe-west1, Belgium) operator: type: string description: >- Legal entity operating the inference service for this deployment, e.g. `Microsoft Corporation`, `Google LLC`, `OpenAI`, `Scaleway SAS`, or `unknown` where unrecorded. example: Google LLC quantization: type: string description: >- Weight precision served. Known values: `bf16`, `fp16`, `fp8`, `fp4`, `nvfp4`, `int4`, `unknown`. example: unknown precisionSource: type: string description: >- How `quantization` is evidenced. Known values: `contractual`, `published`, `reported`, `unknown`. example: unknown certifications: type: array items: type: string description: >- Certifications evidenced for this deployment. Empty when none is recorded. dataRetention: anyOf: - type: string - type: object properties: days: type: integer minimum: 0 required: - days description: >- Prompt retention on this deployment. Known values: `zero`, `metadata-only`, `unknown`, or `{ "days": N }`. example: unknown supportedProviderTools: type: array items: type: string description: >- Provider-executed agent tools this deployment supports. example: - web_search pricing: type: - object - 'null' properties: inputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Input-token price, USD per 1M tokens. Use `inputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md outputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Output-token price, USD per 1M tokens. Use `outputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreationInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Default-retention cache-write input-token price, USD per 1M tokens. Use `cacheCreationInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheCreation1hInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. One-hour cache-write input-token price, USD per 1M tokens. Use `cacheCreation1hInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md cacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Cache-read input-token price, USD per 1M tokens. Use `cacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio input-token price, USD per 1M tokens. Use `audioInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioOutputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio output-token price, USD per 1M tokens. Use `audioOutputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md audioCacheReadInputUsdPer1M: type: number minimum: 0 description: >- Deprecated. Audio cache-read token price, USD per 1M tokens. Use `audioCacheReadInputEurPer1M`. deprecated: true x-deprecation: register_id: CR-model-catalog-usd-pricing deprecated_at: '2026-09-27' replacement: >- The EUR twin on the same object (`inputEurPer1M`, `cacheReadInputEurPer1M`, …; `input_eur_per_1m`, … on MCP). Converted from USD at the `fx` rate each response states, except provider cost on a host that invoices in EUR, which is the published EUR list of that host. rollout_state: announced old_semantics: >- Customer price (after markup) or provider cost (before markup) in USD per 1M tokens, as unrounded numbers. Each is present exactly when the model bills that dimension separately, and keeps that meaning while this entry is open. documentation: >- https://github.com/veryfront/veryfront-api/blob/main/docs/reference/api-surface-inventory.md inputEurPer1M: type: number minimum: 0 description: >- Input-token price. EUR per 1M tokens; see `fx`. outputEurPer1M: type: number minimum: 0 description: >- Output-token price. EUR per 1M tokens; see `fx`. cacheCreationInputEurPer1M: type: number minimum: 0 description: >- Default-retention cache-write input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. cacheCreation1hInputEurPer1M: type: number minimum: 0 description: >- One-hour cache-write input-token price, when the provider bills it apart from the default cache write. EUR per 1M tokens; see `fx`. cacheReadInputEurPer1M: type: number minimum: 0 description: >- Cache-read input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioInputEurPer1M: type: number minimum: 0 description: >- Audio input-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioOutputEurPer1M: type: number minimum: 0 description: >- Audio output-token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. audioCacheReadInputEurPer1M: type: number minimum: 0 description: >- Audio cache-read token price, when the provider bills it separately. EUR per 1M tokens; see `fx`. required: - inputUsdPer1M - outputUsdPer1M - inputEurPer1M - outputEurPer1M description: >- Customer price for this deployment after markup, or null when no price is recorded. Excludes provider cost and the markup amount. required: - tier - jurisdiction - region - processing - processing_label - operator - quantization - precisionSource - certifications - dataRetention - supportedProviderTools - pricing description: >- A model deployment, including its location, terms, and pricing. description: >- Available deployments in routing preference order. Without a policy, requests use the first deployment. Only eligible hosts are listed. If a production deployment is eligible, only production deployments appear. Otherwise, an eligible rollout fallback appears with its own tier, region, and retention settings. required: - id - modelId - provider - providerLabel - providerLogoKey - name - aliases - capabilities - temperature - supportedProviderTools - pricing - providerPricing - pricingSources - deployments description: Available AI models. providers: type: array items: type: string description: Available provider names. defaultModelId: type: string description: Default unified model ID for managed execution. example: mistral/mistral-small-2503 studioDefaultModelId: type: string description: >- Default unified model ID for Studio model selection, subject to the served catalog. example: anthropic/claude-sonnet-4-6 fx: type: object properties: base: type: string description: The currency one unit of which `rate` prices example: EUR quote: type: string description: The currency `rate` is expressed in example: USD rate: type: number exclusiveMinimum: 0 description: 1 `base` = `rate` `quote` example: 1.1403 source: type: string description: Where the rate is published example: ECB euro foreign exchange reference rate date: type: string description: The reference date of the rate (YYYY-MM-DD) example: '2026-09-25' required: - base - quote - rate - source - date description: >- 1 `base` = `rate` `quote`. EUR prices converted from USD are the USD price divided by `rate`, rounded to 6 decimal places. For a host that invoices in EUR the EUR prices are that host's published EUR list, unconverted and unrounded. web_tools: type: array items: type: object properties: tool: type: string description: >- The platform agent tool: `web_search` or `web_fetch`. example: web_search data_sent: type: string description: >- What the tool sends out. Known values: `query_and_search_options` (the search query, plus any domain filters and approximate user location the call states), `url_and_fetched_content` (the URL is requested and the page is returned into the run). An open string: tolerate values you have not seen. example: query_and_search_options recipient_class: type: string description: >- Who receives it. Known values: `third_party_search_provider` (the search backend of the selected deployment), `fetch_target` (the site the URL names). example: third_party_search_provider boundary_covered: type: boolean description: >- Whether this flow is covered by the accepted processing boundary. False unless a deployment carries evidence for it; never inferred from where inference runs. available: type: boolean description: >- Whether agent runs are offered the tool in this environment, under this project's policy. unavailable_reason: type: string description: >- Why the tool is not offered. Known values: `no_search_deployment`, `boundary_not_covered`, `policy_denied`. example: boundary_not_covered required: - tool - data_sent - recipient_class - boundary_covered - available description: >- What a platform web tool sends, to whom, and whether it is offered. description: >- The platform web tools' data flows. Facts about the tool, not about the model: inference residency says nothing about where a search query or fetched URL goes. required: - models - providers - defaultModelId - studioDefaultModelId - fx - web_tools '403': description: Forbidden content: application/problem+json: schema: $ref: '#/components/schemas/Problem' '503': description: This project's inference policy could not be read content: application/json: schema: $ref: '#/components/schemas/AiGatewayError' components: schemas: Problem: type: object properties: cause: type: string enum: - schedule_concurrency_forbidden - schedule_fire_in_progress - schedule_run_limit_reached description: Schedule admission refusal. idempotency: type: object properties: state: type: string enum: - in_progress - payload_mismatch - outcome_unknown operation: type: string minLength: 1 retryable: type: boolean retry_after_seconds: type: - integer - 'null' minimum: 0 expires_at: type: - string - 'null' recovery: type: object properties: available: type: boolean description: type: string resource_url: type: string required: - available - description required: - state - operation - retryable - retry_after_seconds - expires_at description: Replay status, retry timing, and recovery lookup. type: type: string description: URI reference identifying the problem type example: https://api.veryfront.com/errors/validation-failed title: type: string description: Short human-readable summary of the problem example: Validation Failed status: type: integer minimum: 100 maximum: 599 description: HTTP status code example: 400 code: type: string description: Stable machine-readable error code example: RUN_INVALID_INPUT detail: type: string description: Human-readable explanation specific to this occurrence example: The 'email' field must be a valid email address instance: type: string description: URI reference identifying the specific occurrence example: /projects/my-project/members slug: type: string description: Stable identifier for the problem type example: validation-failed category: type: string enum: - AUTH - RESOURCE - VALIDATION - DATABASE - SUBSCRIPTION - INTEGRATION - INTERNAL description: API error category example: VALIDATION suggestion: type: string description: Suggested action to resolve the problem example: >- Check the request body and query parameters against the API documentation. errors: type: array items: type: object properties: field: type: string description: Field path that caused the error example: email message: type: string description: Error message for this field example: Invalid email format code: type: string description: Error code for this field example: INVALID_FORMAT required: - field - message description: Detailed field-level errors for validation failures required: - type - title - status AiGatewayError: type: object properties: error: type: string description: Error message. code: type: string description: Stable machine-readable error code. model: type: string maxLength: 256 description: Requested model ID, on model-specific refusals. successor: type: string description: The model that replaces a retired one, on `model_retired` refusals. retired_on: type: string description: >- UTC date (YYYY-MM-DD) the model was retired, on `model_retired` refusals. allowed_models: type: array items: type: string description: >- The models the calling API key may use, on `key-model-denied` refusals. required: - error securitySchemes: bearerAuth: type: http scheme: bearer bearerFormat: JWT or API Key description: >- Use a JWT bearer token or a Veryfront API key in the `Authorization` header. apiKeyAuth: type: apiKey in: header name: X-API-Key description: Alternative API key header for `vf__` tokens. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.