specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Lamini
providerId: lamini
created: '2026-06-20'
modified: '2026-06-20'
reconciled: true
tags:
- AI
- LLM
- Fine-Tuning
- Memory Tuning
- Inference
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Lamini Platform spend. On the On-Demand tier, Lamini bills a
  flat per-token inference rate and a per-tuning-step rate, drawing down prepaid
  or free credit. Enterprise spend is a negotiated contract for reserved or
  on-premises GPU capacity. The two primary cost drivers are inference token
  volume and tuning steps (which scale with dataset size, epochs, and GPU count).
notes: >-
  On-Demand rates and free-credit amounts change; verify against the Lamini
  pricing page during reconciliation. Enterprise costs are contract-based and
  not metered the same way as On-Demand.
sources:
- https://www.lamini.ai/pricing
- https://www.lamini.ai/blog/lamini-on-demand-300-in-free-credit
- https://docs.lamini.ai
- 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: Lamini
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Lamini Platform
  ServiceCategory: AI and Machine Learning
  ProviderName: Lamini
  PublisherName: Lamini
  InvoiceIssuerName: Lamini
  BillingCurrency: USD
  ChargeCategory: Usage
  PricingCategory: Usage-Based
meters:
- name: inference_tokens
  description: Tokens generated and consumed via /v1/completions, billed at a flat per-1M-token rate.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
  - api
- name: tuning_steps
  description: Tuning steps executed by /v1/train jobs, billed per step and scaling with dataset size, epochs, and GPU count.
  unit: steps
  aggregation: sum
  dimensions:
  - account
  - model
  - job
- name: embedding_requests
  description: Embedding generation requests via /v1/embedding.
  unit: requests
  aggregation: sum
  dimensions:
  - account
  - api
- name: classification_requests
  description: Classification and prediction requests via the /v1/classifier endpoints.
  unit: requests
  aggregation: sum
  dimensions:
  - account
  - model
- name: free_credit
  description: Free credit granted to new accounts, drawn down before paid usage.
  unit: usd
  aggregation: sum
  dimensions:
  - account
principles:
- name: Visibility
  description: Track On-Demand credit burn and inference-token / tuning-step usage per project and API key.
- name: Allocation
  description: Tag API keys and tuning jobs per workload/team; map to internal cost centers.
- name: Optimization
  description: Use Memory Tuning to embed knowledge instead of long prompts, right-size base models, and reserve Enterprise capacity for steady high-volume workloads.
- name: Accountability
  description: Assign owners per project/key; review token and tuning-step spend against budget and free-credit balance.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com