specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Fireworks AI
providerId: fireworks-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
- Multimodal
- Fine-tuning
- GPU
- FinOps
- Cost Management
- FOCUS
description: >-
  FinOps view of Fireworks AI spend. Postpaid usage-based billing in USD,
  with serverless per-token rates per model, embedding rates by model size,
  per-1M-token training-token fine-tuning rates, per-GPU-hour on-demand
  deployments, and a 50% discount on Batch and Cached Input tokens.
notes: >-
  $1 free credit at signup; postpaid invoicing thereafter. Fine-tuned model
  serving costs the same as the base model.
sources:
- https://fireworks.ai/pricing
- https://docs.fireworks.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: Fireworks AI
serviceCategory: AI and Machine Learning
billingModel:
  pricingCategory: Usage-Based
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
  - Usage
  - Purchase
  - Adjustment
focusColumns:
  ServiceName: Fireworks AI
  ServiceCategory: AI and Machine Learning
  ProviderName: Fireworks AI
  PublisherName: Fireworks AI
  InvoiceIssuerName: Fireworks AI
  BillingCurrency: USD
  ChargeCategory: Usage
  PricingCategory: Usage-Based
meters:
- name: input_tokens
  description: Tokens sent in chat / vision requests, billed per 1M tokens per model.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
  - api
- name: cached_input_tokens
  description: Cached-input tokens at 50% of standard input rate.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
- name: output_tokens
  description: Tokens generated by the model, per 1M tokens per model.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
  - api
- name: embedding_tokens
  description: Tokens processed for embeddings, per 1M tokens per embedding-model size class.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
- name: rerank_documents
  description: Documents reranked, per model.
  unit: documents
  aggregation: sum
  dimensions:
  - account
  - model
- name: image_generations
  description: Image generation requests per model and resolution.
  unit: images
  aggregation: sum
  dimensions:
  - account
  - model
- name: audio_seconds
  description: Audio seconds for STT/TTS workloads per model.
  unit: seconds
  aggregation: sum
  dimensions:
  - account
  - model
- name: batch_tokens
  description: Tokens consumed via the Batch API at 50% discount.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - model
- name: fine_tuning_tokens
  description: Training tokens for SFT (LoRA / full) jobs per base-model size class.
  unit: tokens
  aggregation: sum
  dimensions:
  - account
  - base_model
  - method
- name: gpu_seconds
  description: On-demand dedicated GPU runtime per GPU class (H100, H200, B200, B300).
  unit: seconds
  aggregation: sum
  dimensions:
  - account
  - deployment
  - gpu_class
principles:
- name: Visibility
  description: Pull token / GPU-second usage from the Fireworks console; reconcile against postpaid invoices.
- name: Allocation
  description: Tag API keys per workload/team and join usage to internal cost centers.
- name: Optimization
  description: Use Batch (-50%) for non-realtime work; enable prompt caching (-50%); pick smaller embedding models when sufficient; right-size GPU deployments and autoscale to zero where possible.
- name: Accountability
  description: Assign owners per deployment and per fine-tuning job; review monthly token/GPU burn vs. budget.
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com