name: BentoML Plans and Pricing
description: >
  BentoCloud pricing tiers for the managed AI inference platform. The open-source BentoML
  framework is free under Apache 2.0. BentoCloud is billed based on actual compute consumed,
  metered per second with no charges for deployments scaled to zero.
specificationVersion: "0.1"
url: https://www.bentoml.com/pricing
plans:
  - name: Starter
    description: >
      For small teams of developers who want to focus on building AI applications without
      managing infrastructure. Billed monthly based on total usage via credit card.
    price: Pay-as-you-go
    billingCycle: monthly
    features:
      - BentoCloud managed infrastructure
      - Autoscaling with scale-to-zero
      - GPU and CPU instance access
      - BentoCloud Deployment API
      - API token management
      - $10 in free trial credits on signup
      - Access to monitoring and observability
    limits:
      - type: billing
        description: Credit card billing; pay for actual compute consumed, billed per second

  - name: Scale
    description: >
      For teams requiring formal SLAs, cold-start guarantees, and uptime targets.
      Includes all Starter plan features with enhanced support commitments.
    price: Contact Sales
    billingCycle: invoice
    features:
      - All Starter plan features
      - Formal SLA with uptime targets
      - Cold-start guarantees
      - Priority support
      - Invoice-based billing

  - name: Enterprise
    description: >
      For teams that want to use BentoCloud in their own cloud or on-premises environment
      (BYOC — Bring Your Own Cloud). Tailored for organizations requiring data privacy,
      compliance, and full infrastructure control.
    price: Contact Sales
    billingCycle: invoice
    features:
      - All Scale plan features
      - BYOC (Bring Your Own Cloud) deployment
      - On-premises deployment option
      - Custom SLAs
      - Dedicated support
      - Advanced security and compliance controls
      - Invoice billing with custom terms

computePricing:
  description: >
    BentoCloud meters compute by the second. You only pay for active compute; deployments
    scaled to zero incur no cost. Instance types available include CPU and GPU variants.
  instanceTypes:
    - name: cpu.1
      type: CPU
      description: 1 vCPU CPU instance
    - name: cpu.2
      type: CPU
      description: 2 vCPU CPU instance
    - name: cpu.4
      type: CPU
      description: 4 vCPU CPU instance
    - name: cpu.8
      type: CPU
      description: 8 vCPU CPU instance
    - name: gpu.t4.1
      type: GPU
      description: 1x NVIDIA T4 GPU instance
    - name: gpu.l4.1
      type: GPU
      description: 1x NVIDIA L4 GPU instance
    - name: gpu.a100.1
      type: GPU
      description: 1x NVIDIA A100 GPU instance