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: Nvidia
providerId: nvidia
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: false
tags:
  - GPU
  - AI
  - Machine Learning
  - Computing
  - Graphics
  - Rate Limiting
description: NVIDIA's developer API surface is multi-product. build.nvidia.com hosted NIM endpoints have
  per-account / per-API-key rate limits and free-credit budgets that rotate by promotion; specific RPM/TPM
  numbers are not consistently published. Self-hosted NIM (via AI Enterprise license) has no NVIDIA-side
  rate limits — throughput is bounded by the customer's GPU hardware. NGC downloads are throttled per-IP
  by the catalog CDN.
notes: NVIDIA does not publish consistent numeric rate limits across its developer surfaces. Self-hosted
  deployments have no NVIDIA-imposed throttle; hosted endpoints throttle per-key with values that vary
  by service.
sources:
  - https://build.nvidia.com/
  - https://docs.api.nvidia.com/
  - https://docs.nvidia.com/ngc/
responseCodes:
  unauthorized: 401
  forbidden: 403
  throttled: 429
  serviceUnavailable: 503
limits:
  - name: build.nvidia.com hosted endpoints
    scope: api-key
    metric: varies
    limit: per-key throttle and free-credit budget; specific RPM/TPM not publicly documented
    applies:
      - NVIDIA Developer API
  - name: Self-hosted NIM (AI Enterprise)
    scope: cluster
    metric: requests_per_second
    limit: bounded by customer GPU hardware; no NVIDIA-imposed rate limit
    applies:
      - NVIDIA Developer API
  - name: NGC Catalog downloads
    scope: IP
    metric: varies
    limit: CDN-throttled per-IP; not numerically published
    applies:
      - NVIDIA Developer API
policies:
  - name: API key required
    description: All hosted NIM endpoints (build.nvidia.com, integrate.api.nvidia.com) require an NVIDIA
      developer API key passed via Authorization Bearer header.
  - name: Backoff
    description: Implement exponential backoff with jitter on 429 responses. Honor Retry-After when present.
  - name: Free-credit exhaustion
    description: When free-trial credits on build.nvidia.com are exhausted, requests return an authorization
      error rather than a throttle; obtain an AI Enterprise license or cloud-marketplace credentials for
      production.
  - name: Self-host for high throughput
    description: For sustained high-throughput inference, deploy NIM containers on customer GPU infrastructure
      under AI Enterprise rather than calling hosted endpoints.
maintainers:
  - FN: API Evangelist
    email: info@apievangelist.com