specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Moonshot AI
providerId: moonshot-ai
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
- LLM
- Inference
- Long Context
- Kimi
- FinOps
- Cost Management
- FOCUS
description: >-
  FOCUS-aligned FinOps profile for Moonshot AI / Kimi. Billing model is prepaid recharge with
  pay-as-you-go consumption metered against per-million-token rates per model and per direction
  (input cache-hit, input cache-miss, output). Batch jobs are billed at 50% of online rates.
notes: >-
  Prepaid balance maps to FOCUS Purchase / Adjustment categories; consumption maps to Usage.
  Document extraction and storage are free at reconciliation; expect them to monetize later.
sources:
- https://platform.kimi.ai/docs/pricing/chat
- https://platform.kimi.ai/docs/pricing/batch
- https://platform.kimi.ai/docs/pricing/limits
- 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: Moonshot AI
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Usage Based
  billingFrequency: On-Demand (Recharge)
  billingCurrency: CNY/USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Moonshot AI Platform
  ServiceCategory: AI and Machine Learning
  ProviderName: Moonshot AI
  PublisherName: Moonshot AI
  InvoiceIssuerName: Moonshot AI
  BillingCurrency: CNY
  ChargeCategory: Usage
meters:
- name: input_tokens_cache_miss
  description: Input tokens billed when cache miss occurs.
  unit: tokens
  aggregation: sum
  dimensions:
  - model
  - account
- name: input_tokens_cache_hit
  description: Discounted input tokens served from prompt cache.
  unit: tokens
  aggregation: sum
  dimensions:
  - model
  - account
- name: output_tokens
  description: Generated output tokens.
  unit: tokens
  aggregation: sum
  dimensions:
  - model
  - account
- name: batch_tokens
  description: Batch endpoint tokens (50% discount versus online).
  unit: tokens
  aggregation: sum
  dimensions:
  - model
  - account
- name: web_search_calls
  description: Web search tool invocations.
  unit: calls
  aggregation: sum
  dimensions:
  - account
principles:
- name: Visibility
  description: Use the Balance API and console usage exports to track recharge balance and consumption.
- name: Allocation
  description: Tag API keys per workload/team to attribute model spend.
- name: Optimization
  description: Prefer prompt caching, batch endpoints (50% off), shorter contexts, and right-sized models.
- name: Accountability
  description: Monitor tier ceiling utilization and renegotiate at $3K/$10K cumulative-recharge tier breaks.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com