specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Replicate
providerId: replicate
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: '2026-05-04'
reconciled: true
tags:
  - FinOps
  - FOCUS
  - ML Inference
description: FOCUS-aligned FinOps for Replicate.
sources:
  - https://replicate.com/pricing
alignedWith:
  framework: FinOps Foundation Framework
  frameworkUrl: https://www.finops.org/framework/
  dataSpec: FOCUS
  dataSpecVersion: '1.3'
  dataSpecUrl: https://focus.finops.org/focus-specification/v1-3/
publisherName: Replicate
serviceCategory: ML Inference
billingModel:
  pricingCategory: Per-Second GPU
  billingFrequency: Monthly
  billingCurrency: USD
focusColumns:
  ServiceName: Replicate
  ServiceCategory: ML Inference
  ProviderName: Replicate
  PublisherName: Replicate
  BillingCurrency: USD
meters:
  - name: gpu_seconds
    unit: second
    aggregation: sum
    dimensions:
      - hardware
      - model
  - name: predictions_count
    unit: prediction
    aggregation: sum
  - name: deployment_idle_seconds
    unit: second
    aggregation: sum
principles:
  - name: Visibility
    description: Track Replicate consumption monthly via admin/billing exports.
  - name: Allocation
    description: Tag usage to teams or cost centers for chargeback.
  - name: Optimization
    description: Right-size models/tiers; use caching and batching.
  - name: Accountability
    description: Set hard spend limits; quarterly review.