specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: SUTRA (Two AI)
providerId: sutra-ai
created: '2026-06-21'
modified: '2026-06-21'
reconciled: false
tags:
- AI
- LLM
- Multilingual
- Inference
- Reasoning
- Rate Limiting
- Quotas
- Throttling
description: >-
  The SUTRA API is OpenAI-compatible and enforces per-account limits on
  inference. Because the public documentation directs developers to contact
  Two AI (Numeric) for API access, specific per-account RPM (requests per
  minute) and TPM (tokens per minute) limits are not published and are not
  reconciled in this artifact. As with other OpenAI-compatible services,
  expect HTTP 429 responses when limits are exceeded.
notes: >-
  Confirm per-account and per-model RPM / TPM ceilings directly with Two AI on
  reconciliation; limits likely vary by plan, startup-program status, and
  enterprise agreement.
sources:
- https://docs.two.ai/docs/getting-started
- https://docs.two.ai/docs/models/sutra-v2-guide
responseCodes:
  throttled: 429
limits:
- name: Requests Per Minute (RPM)
  scope: account
  metric: requests
  limit: see provider (not published)
  notes: Per-account RPM; not publicly documented.
- name: Tokens Per Minute (TPM)
  scope: account
  metric: tokens
  limit: see provider (not published)
  notes: Per-account TPM; not publicly documented.
- name: Concurrent Requests
  scope: account
  metric: requests
  limit: see provider (not published)
  notes: Concurrency ceiling varies by plan/agreement.
policies:
- name: Tiered Limits
  description: Limits vary by plan, startup-program status, and enterprise agreements.
- name: Backoff Strategy
  description: Clients should implement exponential backoff with jitter and honor Retry-After on 429 responses.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com