specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: fal
providerId: fal-ai
created: '2026-05-25'
modified: '2026-05-25'
reconciled: true
tags:
  - FinOps
  - Cost Management
  - FOCUS
  - AI
  - GPU
  - Inference
description: 'FOCUS-aligned FinOps definition for fal — generative media inference. Three distinct billing lanes:
  per-output Model APIs, per-second/per-hour Serverless and Compute GPU rentals, and stored asset egress from the
  fal CDN.'
sources:
  - https://fal.ai/pricing
  - https://fal.ai/docs
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: fal (Features and Labels, Inc.)
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
    - Usage
    - Purchase
    - Credit
    - Adjustment
  primaryUnit: output
  secondaryUnits:
    - megapixel
    - video_second
    - audio_minute
    - gpu_second
    - gpu_hour
    - cdn_gb
focusColumns:
  ServiceName: fal Model APIs
  ServiceCategory: AI and Machine Learning
  ProviderName: fal
  PublisherName: fal
  InvoiceIssuerName: Features and Labels, Inc.
  PricingCategory: Usage-Based
  PricingUnit: output
  BillingCurrency: USD
meters:
  - name: image_outputs
    unit: image
    aggregation: sum
    dimensions:
      - model_id
      - api_key
      - timestamp
  - name: image_megapixels
    unit: megapixel
    aggregation: sum
    dimensions:
      - model_id
      - api_key
  - name: video_seconds
    unit: video_second
    aggregation: sum
    dimensions:
      - model_id
      - api_key
      - aspect_ratio
  - name: audio_minutes
    unit: audio_minute
    aggregation: sum
    dimensions:
      - model_id
      - api_key
  - name: gpu_seconds
    unit: gpu_second
    aggregation: sum
    dimensions:
      - gpu_class
      - app_id
      - instance_id
  - name: gpu_hours
    unit: gpu_hour
    aggregation: sum
    dimensions:
      - gpu_class
      - instance_id
  - name: cdn_egress_gb
    unit: cdn_gb
    aggregation: sum
    dimensions:
      - file_id
discounts:
  - name: Volume / Enterprise
    description: Negotiated discounts on Model API output prices and reserved GPU capacity for enterprise contracts.
  - name: Day-zero promo credits
    description: Promotional credits issued at launch for newly hosted day-zero models.
exports:
  - format: CSV
    surface: Dashboard usage export
  - format: JSON
    surface: GET /billing/usage (programmatic)
  - format: PDF
    surface: GET /billing/invoices (monthly statements)
notes:
  - Model APIs are billed per output (per image, per video-second, per audio-minute, or per megapixel) depending
    on model; the unit appears in the fal.ai/pricing table.
  - Serverless apps and dedicated Compute instances are billed per GPU-second (Serverless) and per GPU-hour or
    GPU-second (Compute) — H100 / H200 / A100 / B200 classes.
  - CDN delivery from v3.fal.media is included with model output today; egress accounting is being introduced
    for high-volume serving.