Modal

Serverless cloud for AI and ML workloads

Free + usage | Team $250/mo + usageComputeLast verified: 2026-02-02

Overview

Modal is a serverless compute platform that lets you run AI workloads without managing infrastructure. Pay only for actual compute time by the CPU cycle, with no charges for idle resources. Access GPUs from T4 to H200 with per-second billing. Containers cold start in milliseconds, and the platform autoscales automatically based on demand. Includes $30/month free credits on the Starter plan. SOC 2 compliant with support for AWS and GCP marketplace committed spend. Ideal for ML training, batch inference, data processing, and deploying AI applications. Startups can apply for up to $25k in credits, academics up to $10k.

Works with

REST API

Pricing

$0 + computeStarter

$30/mo free credits
3 workspace seats
100 containers
10 GPU concurrency
Region selection

$250/mo + computeTeam

$100/mo free credits
Unlimited seats
1000 containers
50 GPU concurrency
Custom domains
Deployment rollbacks

CustomEnterprise

Volume discounts
Higher GPU concurrency
HIPAA compliance
SSO and audit logs
Dedicated support

Pros

+True serverless (no idle costs)
+Per-second GPU billing
+$30/mo free credits
+Fast container cold starts

Cons

-Learning curve for serverless model
-Region selection costs extra

Get Discovered by Developers

Promote your tool

Reach thousands of developers actively searching for AI tools. Featured listings get 10x more clicks.

Back