specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Adept
providerId: adept
created: '2026-05-08'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-08, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-08'
reconciled: true
tags:
- AI
- Agents
- Foundation Models
- Action Models
- Open Source
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Adept "spend." Adept does not invoice customers - it has
  no commercial API. The only direct cost surface is downloading the open-
  source Fuyu-8B and Persimmon-8B models from Hugging Face (free) and the
  customer's own self-hosted GPU inference cost (or third-party inference
  provider cost). Those costs appear under the inference provider's
  ServiceName, not Adept's.
notes: >-
  In a corporate FOCUS dataset there is typically no Adept-issued invoice
  line item. Track the Adept models via tagging on the inference provider's
  spend (e.g., AWS / Together / Fireworks running Fuyu-8B).
sources:
- https://www.adept.ai/
- https://huggingface.co/adept
- 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: Adept
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Other
  billingFrequency: One-Time
  billingCurrency: USD
  chargeCategories:
  - Adjustment
focusColumns:
  ServiceName: Adept Open-Source Models
  ServiceCategory: AI and Machine Learning
  ProviderName: Adept
  PublisherName: Adept
  InvoiceIssuerName: Not Applicable
  BillingCurrency: USD
  ChargeCategory: Adjustment
  PricingCategory: Other
meters:
- name: model_downloads
  description: Adept open-source model downloads (free) used for cost-free attribution / inventory tracking only.
  unit: downloads
  aggregation: sum
  dimensions:
  - team
  - model
- name: self_hosted_inference_cost
  description: Customer's own GPU inference cost when serving Adept open-source models (recorded under the inference provider, not Adept).
  unit: usd
  aggregation: sum
  dimensions:
  - team
  - model
  - inference_provider
principles:
- name: Visibility
  description: Inventory which teams downloaded Adept models; correlate with self-hosted GPU spend tagged to those models.
- name: Allocation
  description: Tag GPU inference workloads with the model name (Fuyu-8B / Persimmon-8B) for chargeback.
- name: Optimization
  description: Quantize or distill Adept open-source models if hosting cost exceeds value; consider replacing with newer open-source action models.
- name: Accountability
  description: No Adept invoice to reconcile; owners are the teams running self-hosted inference.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com