specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Martian
providerId: martian-ai
created: '2026-06-20'
modified: '2026-06-20'
reconciled: false
tags:
- AI
- LLM
- Model Router
- Gateway
- Cost Optimization
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Martian Gateway spend. Martian is an LLM model router whose
  core value proposition is cost optimization: for each request it routes to
  the model that meets quality targets at the lowest cost, with controls such
  as willingness-to-pay to trade quality against price. The marginal cost of a
  routed request is primarily the pass-through cost of the selected upstream
  provider model; Martian's own gateway unit rates are not publicly enumerated
  and are not reconciled in this artifact.
notes: >-
  Verify Martian gateway markups, free-allotment terms, and any per-request
  fee against Martian during reconciliation. Because routing decisions change
  which upstream model serves a request, per-request cost varies with the
  selected model and the configured cost/quality trade-off.
sources:
- https://www.withmartian.com
- https://docs.withmartian.com/quickstart
- 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: Martian
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Martian Gateway
  ServiceCategory: AI and Machine Learning
  ProviderName: Martian
  PublisherName: Martian
  InvoiceIssuerName: Martian
  BillingCurrency: USD
  ChargeCategory: Usage
  PricingCategory: Usage-Based
meters:
- name: routed_requests
  description: Requests routed through the Martian Gateway, metered beyond the free allotment.
  unit: requests
  aggregation: sum
  dimensions:
  - account
  - selected_model
- name: input_tokens
  description: Tokens sent in routed chat / messages requests; cost depends on the selected upstream model.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - selected_model
- name: output_tokens
  description: Tokens generated by the selected upstream model on a routed request.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - selected_model
principles:
- name: Visibility
  description: Inspect Martian usage and per-request routing decisions to see which upstream models are serving spend.
- name: Allocation
  description: Tag API keys per workload/team and map routed spend to internal cost centers.
- name: Optimization
  description: Tune the willingness-to-pay / cost-quality trade-off so the router selects cheaper models when sufficient; rely on automatic failover to avoid costly retries.
- name: Accountability
  description: Assign owners per project/key and review routed spend monthly against budget.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com