> ## Documentation Index > Fetch the complete documentation index at: https://docs.runanywhere.ai/llms.txt > Use this file to discover all available pages before exploring further. # List models available to the current key ## OpenAPI ````yaml /api-reference/openapi.json get /models openapi: 3.1.0 info: description: >- OpenAI-compatible hosted inference for Wally. Usage is billed against prepaid credits at the list prices shown in the RunAnywhere console. title: RunAnywhere Wally Public Inference API version: 2.0.0 servers: - description: Wally production inference gateway url: https://inference.runanywhere.ai/v1 security: - BearerAuth: [] paths: /models: get: summary: List models available to the current key operationId: listModels responses: '200': content: application/json: schema: $ref: '#/components/schemas/ModelList' description: Available model aliases '401': $ref: '#/components/responses/Unauthorized' '429': $ref: '#/components/responses/RateLimited' default: $ref: '#/components/responses/Error' components: schemas: ModelList: additionalProperties: false properties: data: items: $ref: '#/components/schemas/PublicModel' type: array object: const: list required: - object - data type: object PublicModel: additionalProperties: false properties: created: description: >- OpenAI-shaped creation timestamp (unix seconds) the gateway reports; informational. minimum: 0 type: integer id: $ref: '#/components/schemas/PublicModelId' max_input_tokens: description: >- The advertised ceiling on input tokens for this model -- what an OpenAI-shaped coding agent reads to decide when to compact. It is the largest prompt the engine still answers with one output token, not the raw context length: a prompt longer than this is refused with 400 before it is billed. null means the deployment has not established one. minimum: 1 type: - integer - 'null' max_output_tokens: description: >- The advertised ceiling on output tokens, when the deployment declares one separately from the context window. null means this model draws output from the single budget max_input_tokens already describes. minimum: 1 type: - integer - 'null' mode: description: The gateway's model mode; every Wally model is a chat model. enum: - chat type: string object: const: model owned_by: $ref: '#/components/schemas/ModelOwner' required: - id - object - owned_by type: object OpenAIError: additionalProperties: false properties: error: $ref: '#/components/schemas/OpenAIErrorDetail' required: - error type: object RequestId: maxLength: 256 minLength: 1 type: string AccessControlExposeHeaders: description: >- The response headers a browser client may read, as a comma-separated list; it names x-request-id, Retry-After and Server-Timing when the response carries them. maxLength: 1024 minLength: 1 type: string RetryAfterSeconds: minimum: 0 type: integer ServerTiming: description: >- W3C Server-Timing. As a HEADER it carries pre-stream phases only, for example the gateway's authorize call and the sidecar's ra-queue (time waiting for a seat) and ra-admit (the sidecar's own work before the request reached the model). Time to first token is never in the header, because it is not known when the headers are sent; on a streaming response it arrives as a Server-Timing TRAILER named ra-ttft, after the last SSE frame. Metric names are not a closed set: ignore a name you do not know, and treat the trailer as optional. maxLength: 1024 minLength: 1 type: string PublicModelId: maxLength: 128 minLength: 1 pattern: ^[A-Za-z0-9][A-Za-z0-9._:/-]*$ type: string ModelOwner: maxLength: 128 minLength: 1 type: string OpenAIErrorDetail: additionalProperties: false properties: code: anyOf: - $ref: '#/components/schemas/ErrorCode' - type: 'null' message: $ref: '#/components/schemas/ErrorMessage' param: anyOf: - $ref: '#/components/schemas/ErrorParameter' - type: 'null' type: $ref: '#/components/schemas/ErrorType' required: - message - type type: object ErrorCode: enum: - bad_request - not_found - payload_too_large - unprocessable - invalid_api_key - expired_api_key - revoked_api_key - insufficient_credit - account_frozen - model_not_entitled - not_in_cohort - user_rate_limited - request_cancelled - quota_exceeded - capacity_exceeded - idempotency_key_reused - unsupported_endpoint - proxy_retired - gateway_unavailable - control_plane_overloaded - credit_gate_unavailable - upstream_contract_violation - unattributed_key - internal - bad_gateway - draining - spool_unhealthy - cold - timeout type: string ErrorMessage: maxLength: 2048 minLength: 1 type: string ErrorParameter: maxLength: 256 minLength: 1 type: string ErrorType: enum: - invalid_request_error - authentication_error - permission_error - rate_limit_error - quota - capacity - server_error - timeout type: string responses: Unauthorized: content: application/json: schema: $ref: '#/components/schemas/OpenAIError' description: Missing, malformed, or unknown Cloud key headers: x-request-id: description: Request correlation id schema: $ref: '#/components/schemas/RequestId' RateLimited: content: application/json: schema: $ref: '#/components/schemas/OpenAIError' description: Entitlement or GPU admission capacity limit headers: Access-Control-Expose-Headers: description: Response headers a browser client may read schema: $ref: '#/components/schemas/AccessControlExposeHeaders' Retry-After: description: Seconds before retrying when supplied schema: $ref: '#/components/schemas/RetryAfterSeconds' Server-Timing: description: Pre-stream phase timings when supplied schema: $ref: '#/components/schemas/ServerTiming' x-request-id: description: Request correlation id schema: $ref: '#/components/schemas/RequestId' Error: content: application/json: schema: $ref: '#/components/schemas/OpenAIError' description: OpenAI-shaped error; raw upstream bodies are never returned headers: x-request-id: description: Request correlation id schema: $ref: '#/components/schemas/RequestId' securitySchemes: BearerAuth: bearerFormat: RunAnywhereCloudKey scheme: bearer type: http ````