generated: '2026-07-21'
method: searched
source: https://docs.sambanova.ai/cloud/docs/features/openai-compatibility, https://docs.sambanova.ai/cloud/docs/models/rate-limits, https://docs.sambanova.ai/cloud/docs/api-reference/using-the-api/api-error-codes
authentication:
  style: bearer-api-key
  header: 'Authorization: Bearer <key>'
  alt_header: 'x-api-key: <key>  (Messages API routes, for Anthropic SDK compatibility)'
  base_url: https://api.sambanova.ai/v1
  ref: authentication/sambanova-systems-authentication.yml
idempotency:
  supported: false
  note: No idempotency-key header or parameter is documented or present in the OpenAPI; inference calls are non-idempotent by nature.
pagination:
  supported: false
  note: Inference endpoints are single request/response; no list-pagination surface. GET /models returns a full list object.
streaming:
  supported: true
  style: server-sent-events
  param: stream=true
  note: Streaming responses return chunks that may contain multiple tokens; count all tokens per chunk for TPS metrics.
request_tracing:
  field: request_id
  note: Every error response (and completion id) carries a request_id / id for tracing; provide request_id to support when reporting issues.
versioning:
  scheme: uri-path
  current: v1
  ref: lifecycle/sambanova-systems-lifecycle.yml
error_envelope:
  shape: '{ "error": { "message", "type", "param", "code" }, "request_id" }'
  format: openai-error-object
  ref: errors/sambanova-systems-error-codes.yml
rate_limit_signaling:
  headers:
    - x-ratelimit-limit-requests
    - x-ratelimit-remaining-requests
    - x-ratelimit-reset-requests
    - x-ratelimit-limit-requests-day
    - x-ratelimit-remaining-requests-day
    - x-ratelimit-reset-requests-day
  measured_in: [RPM, RPD, TPD]
  ref: https://docs.sambanova.ai/cloud/docs/models/rate-limits
compatibility:
  openai: true
  anthropic_messages: true
  note: SambaNova inference APIs are drop-in compatible with the OpenAI client and the Anthropic Messages API.