Modal
Serverless GPUs and containers for Python: run AI inference, training, and batch jobs. · Modal · Compute
Overview
Modal refills $30 of credits every month on the Starter plan.
Good fit
Use when you need GPUs on demand for Python work.
- AI inference
- batch jobs
Not a fit
Avoid if you need always-on GPUs for free.
- always-on services
Quickstart · 3 steps
- pip install modal.
- Run modal setup to log in.
- Decorate a Python function and run modal run.
Other free compute options
Free tierWhat you get freeRiskCard
Cloudflare · Compute
100,000 requests/day · 10ms CPU time/request · Up to 100 Workers per account
Billing risk: None
No card
Vercel · Compute
1,000,000 Edge Requests/month included · 100 GB Fast Data Transfer/month included · 1 team seat on the Hobby plan
Billing risk: None
No card
Vercel · Compute
1,000,000 Function Invocations/month included · 4 hours Fluid Active CPU/month included · 360 GB-hrs Fluid Provisioned Memory/month included
Billing risk: None
No card
Trigger.dev · Compute
$5 compute credits per month · 20 concurrent runs · Unlimited tasks
Billing risk: None
No card