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: Martian
providerId: martian-ai
created: '2026-06-20'
modified: '2026-06-20'
reconciled: false
tags:
- AI
- LLM
- Model Router
- Gateway
- Cost Optimization
- Rate Limiting
- Quotas
- Throttling
description: >-
  The Martian Gateway is a routing layer in front of many upstream provider
  models, so effective throughput is governed both by Martian account limits
  and by the rate limits of whichever provider model the router selects.
  Martian documents a free request allotment and returns standard HTTP 429
  responses when limits are exceeded, but specific per-account RPM/TPM
  values are not publicly enumerated and are not reconciled here.
notes: >-
  Confirm per-account and per-tier request/token limits with Martian during
  reconciliation. Because Martian routes across providers, downstream
  provider rate limits may also apply to a given request.
sources:
- https://docs.withmartian.com/api-reference/endpoints
- https://docs.withmartian.com/api-reference/errors
- https://www.withmartian.com
responseCodes:
  throttled: 429
limits:
- name: Requests Per Minute (RPM)
  scope: account
  metric: requests
  limit: see provider documentation
  notes: Per-account request rate; not publicly enumerated.
- name: Tokens Per Minute (TPM)
  scope: account
  metric: tokens
  limit: see provider documentation
  notes: Per-account token throughput; not publicly enumerated.
- name: Free Request Allotment
  scope: account
  metric: requests
  limit: 2500
  notes: Documented free developer allotment before metered paid usage applies.
- name: Downstream Provider Limits
  scope: model
  metric: requests
  limit: varies by routed provider
  notes: The selected upstream model's own provider rate limits may constrain a routed request.
policies:
- name: Tiered Limits
  description: Limits raise from the free developer tier to paid and Enterprise agreements.
- name: Backoff Strategy
  description: Clients should implement exponential backoff with jitter and honor Retry-After on 429 responses.
- name: Automatic Failover
  description: Martian can fail over across providers, which can mitigate single-provider throttling for a routed request.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com