Building blocks · by Modal Labs
Modal
Offload heavy Python jobs to cloud graphics chips by the second.
code → link, deployed · code → text, data · code, schedule → deployed
- Way in
- Official APISign up yourself, api key
- To start
- Free tierStarter $0 + $30/mo compute credit
- Runs
- In the cloud
- Checked
- Nov 3, 2026Prices as of that day
What it can do 3
Run Python functions or model servers on per-second GPUsStarter: 100 containers, 10 GPU concurrencycode → link, deployedT4 $0.59/h, L4 $0.80/h, A10 $1.10/h, L40S $1.95/h, A100 80GB $2.50/h, H100 $3.95/h, H200 $4.54/h, B200 $6.25/h (per-second billing)
Run untrusted code in isolated sandboxescode → text, dataCPU $0.0000394/core-s, mem $0.0000067/GiB-s (3x list, non-preemptible) + GPU
Schedule recurring jobs (cron) in the cloudcode, schedule → deployedcompute time only; CPU $0.047/core-h
Does the same job
Prices and limits as of Nov 3, 2026, from Modal Labs’s own pages. They change; check before you rely on them.Not affiliated with Modal Labs. Not advice.Something wrong?