Pricing That
Scales with You

Usage-based pricing at any scale

Pricing

Utilize Union and only pay for the resources managed by the platform. Benefit from spot instances with recovery, enhanced caching, and the ability to scale to zero.

Free
(w/ support options available)

DIY ML orchestration for teams with on-premise or bare metal deployments.

Open source (Apache 2.0)
Battle-tested at scale
Deployable on-premise
Vibrant community
Union support plans available
Union.ai
BYOC – Startup
$500/mo + % of compute
(on-demand retail pricing)

A secure and scalable platform ideal for small teams and early stage companies.

Optimized and expanded Flyte
Up to 5 seats
Single cluster/cloud
Purchase on AWS or GCP marketplace
Standard bring-your-own-cloud (BYOC) deployment
Union.ai
BYOC – Enterprise
Custom
(w/ committed use discounts)

Built for enterprises that require large-scale, highly available, and customizable deployments.

Optimized and expanded Flyte
Unlimited seats
Multi-cluster/cloud
Purchase on AWS or GCP marketplace
Customizable bring-your-own-cloud (BYOC) deployment

Feature comparison

Flyte
Serverless
Startup
Enterprise
Actors (reusable containers)
Coming soon
Multi-cluster
Default single cluster,
Multi-cluster with no isolation
Scalable serverless environment
Default single cluster
Multi-cluster for load-balancing, project/domain isolation
Multi-cloud
 AWS / GCP / Azure (coming soon)
Engine
Open Source
Optimized (up to 10x faster)
Optimized (up to 10x faster)
Optimized (up to 10x faster)
Faster data reads with pre-mounted volumes
Managed compute plugins
Supported Ray, Dask, Spark, distributed training with dashboards
Self-serve secrets
Real-time logs
Interactive debugging in UI
(Secure VS Code)
(Secure VS Code)
(Secure VS Code)
Artifact & model registry
Reactive workflows
(up to 5 active triggers)
(up to 100 active triggers)
(unlimited)
Cloud-based image builder
Extensions
Self-managed
Standard Agents (coming soon)
Standard Agents
Standard & User-Developed Agents
Fractional GPUs, spot instances, TPUs
Self-managed
Scale down to zero when not in use
Always-on control plane
Task-level resource metrics
Cluster-wide observability
Included (per-project resource monitoring)
Included (per-project resource monitoring)
Custom workload to nodepool mapping
Self-managed
Deployed in your cloud
Self-managed
RBAC with custom policies
Globally available
Self-managed
High availability w/ SLAs
Security (SOC-2 Type II)
Authentication
Self-managed
SSO with Github
SSO (Google / Okta)
Customizable SSO
Seats
Self-managed
Limited to 1 seat
Up to 5 seats
Unlimited
VPC management
Self-managed
Union-managed VPC
Union-managed VPC
Union-managed VPC,
self-managed VPC
Retention
Self-managed
1 week
1 month
Custom
Support
Flyte Community
(with optional Union support)
Basic support included
Premier support included
Premier support included,
Priority support available
Solutions engineering
(from the team that built Flyte)
On Prem
Self-managed

*Serverless compute pricing

Sign up with GitHub & instantly run workflows at cloud scale. Get $30 in free compute by signing up.

CPU
CPU Core:
$0.12 / Core-Hr
GPU
T4:
A100 (40GB):
$0.71 / GPU-Hr
$4.48 / GPU-Hr
Memory:
$0.029 / GB-Hr