specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Nscale
providerId: nscale
created: '2026-06-21'
modified: '2026-06-21'
reconciled: false
tags:
- AI
- GPU
- Inference
- Serverless
- Cloud Compute
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Nscale spend. Nscale bills usage-based per-token rates for
  serverless text inference (input and output), per-megapixel rates for image
  generation, and per-GPU-hour rates for compute instances, nodes, and
  clusters. Reserved capacity and enterprise commitments reduce effective
  GPU-hour rates.
notes: >-
  Per-model and per-GPU-hour rates change frequently and were not all
  individually confirmed (reconciled false); verify against the Nscale pricing
  page and console during reconciliation.
sources:
- https://www.nscale.com/pricing
- https://www.nscale.com/product/serverless
- https://www.nscale.com/product/gpu-nodes
- https://focus.finops.org/focus-specification/v1-3/
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: Nscale
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Nscale Serverless Inference and GPU Compute
  ServiceCategory: AI and Machine Learning
  ProviderName: Nscale
  PublisherName: Nscale
  InvoiceIssuerName: Nscale
  BillingCurrency: USD
  ChargeCategory: Usage
  PricingCategory: Usage-Based
meters:
- name: input_tokens
  description: Tokens sent in serverless chat / completions requests, billed per 1M tokens per model.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
  - api
- name: output_tokens
  description: Tokens generated by serverless inference, billed per 1M tokens per model.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
  - api
- name: embedding_tokens
  description: Tokens embedded via the embeddings endpoint, billed per 1M tokens per model.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
- name: image_megapixels
  description: Megapixels of generated imagery, billed per megapixel and scaled by diffusion steps.
  unit: megapixels
  aggregation: sum
  dimensions:
  - account
  - model
- name: gpu_hours
  description: GPU-hours consumed by compute instances, nodes, and clusters, billed per accelerator type.
  unit: gpu_hours
  aggregation: sum
  dimensions:
  - account
  - accelerator
  - region
principles:
- name: Visibility
  description: Pull Nscale usage and billing exports; inspect per-model token burn and per-accelerator GPU-hours.
- name: Allocation
  description: Tag API keys and projects per workload/team; map serverless and compute spend to internal cost centers.
- name: Optimization
  description: Pick smaller open-source serverless models when sufficient; cap diffusion steps and image resolution; use reserved GPU capacity for steady compute workloads.
- name: Accountability
  description: Assign owners per project/key; review token and GPU-hour spend monthly against budget.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com