Serverless compute for running Python code on GPUs and CPUs, billed per second of use, with a free Starter plan that includes $30 of monthly credits.
Python developers who want to run inference, batch jobs or fine-tuning on GPUs billed per second, without managing servers, and who have bursty rather than constant load.
Teams that want a per-token model API with no code to deploy: Modal bills compute time, so you bring and host the model yourself. Also poor fit for workloads that need more than 10 concurrent GPUs on the free plan, or need guaranteed non-preemptible capacity at base price, which costs 3x.
Pros
Cons
Modal bills GPUs per second of use. Listed rates include $0.001097 per second for an H100 SXM5, $0.001261 for an H200 SXM, $0.001736 for a B200 and $0.000164 for a T4. The page shows these as $ without naming a currency code.
Yes. Starter is $0 per month plus compute and includes $30 per month in free credits, 3 workspace seats, 100 containers, 10 GPU concurrency and 1 TiB per month of free egress. You pay for usage beyond the credits.
Team is $250 per month plus compute. It includes $100 per month in credits, unlimited seats, 5,000 containers, 50 GPU concurrency and 10 TiB per month of free egress. Enterprise pricing is custom with volume discounts and 100 TiB egress.
CPU is $0.0000131 per physical core per second, where one core equals 2 vCPUs, with a minimum of 0.125 cores per container. Memory is $0.00000222 per GiB per second. Volumes are $0.09 per GiB per month with 1 TiB per month free.
Yes. Sandbox CPU is listed at $0.00003942 per core per second and memory at $0.00000667 per GiB per second, labelled non-preemptible and 3x list. GPU use in sandboxes is charged at standard GPU pricing.
Facts sourced from the vendor’s own site, last verified 2026-10-30.
The Weekly Source
One email a week. New listings, repriced tools, and the ones we would skip - including this one if its pricing changes.
