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: Hyperbolic
providerId: hyperbolic-ai
created: '2026-05-25'
modified: '2026-05-25'
reconciled: true
tags:
  - AI
  - Inference
  - Rate Limiting
  - Quotas
description: Reconciled rate limits for the Hyperbolic Serverless Inference API. Tier-based RPM ceilings
  applied across Chat Completions, Completions, Image Generation, and Audio Generation. No published
  per-token-per-minute (TPM) limits — Hyperbolic emphasises "zero quota limitations" and falls back to
  RPM and account balance as the primary backpressure mechanisms.
sources:
  - https://docs.hyperbolic.ai/inference/overview
  - https://docs.hyperbolic.ai/inference/quickstart
  - https://www.hyperbolic.ai/
algorithm: request-rate
responseCodes:
  throttled: 429
  quotaExceeded: 429
  paymentRequired: 402
spendLimits:
  - tier: Basic
    creditPurchase: $0
    monthlyCap: account balance
  - tier: Pro
    creditPurchase: $5 minimum
    monthlyCap: account balance
  - tier: Enterprise
    creditPurchase: contract
    monthlyCap: contract
limits:
  - tier: Basic
    surface: All inference endpoints
    rpm: 60
    notes: Free tier, community support, no deposit required.
  - tier: Pro
    surface: All inference endpoints
    rpm: 600
    notes: Requires $5+ deposit. Priority and email support.
  - tier: Enterprise
    surface: All inference endpoints
    rpm: -1
    notes: Unlimited via custom enterprise contract. 24/7 dedicated support.
gpuRental:
  notes: GPU marketplace rentals are capacity-bounded rather than rate-limited. Provisioning is rate-limited
    implicitly by available marketplace supply at any moment.