Modal

What is Modal?

Modal is a serverless cloud platform for running AI and data workloads from Python code, with GPUs and CPUs that scale up and down automatically. Teams use it for model inference, training and fine-tuning, high-performance batch jobs and, through its Sandboxes product, isolated environments for AI agents and reinforcement learning rollouts. You define the container image and resources in code, and Modal bills for the time your containers load and process inputs, with no charge once an app has scaled to zero. Pricing is pay-per-use: GPU rates run from $0.000164 per second for a T4 to $0.001972 per second for a B300, plus CPU and memory rates. The Starter plan costs $0 plus compute and includes $30 per month of compute, the Team plan is $250 per month plus compute and includes $100, and Enterprise is custom. Startups and academic researchers can apply for free compute credits.

Use cases

• Serve model inference on GPUs
• Run batch jobs in parallel
• Fine-tune and train models
• Run untrusted agent code in sandboxes
• Schedule cron jobs in the cloud
• Run reinforcement learning rollouts

Who it's for

• ML engineers
• AI startups
• Developers
• Researchers
• Platform teams

Why use Modal

• You pay only for compute actually used, so spiky or unpredictable inference traffic does not require reserved GPUs.
• Containers spin up in 1-2 seconds and scale back to zero, which keeps idle cost low.
• Infrastructure is defined in code, so images, GPUs and schedules live with your application.
• Sandboxes give agents and RL jobs isolated environments that can run in large parallel numbers.
• Credit grants for startups and academics lower the cost of getting started.

Modal features

• Per-second billing for GPUs, CPUs and memory
• GPU choices from Nvidia T4 up to B300
• Autoscaling to zero with containers starting in 1-2 seconds
• Sandboxes for agents and reinforcement learning, with egress controls
• Custom images defined in code
• Cron jobs, distributed queues and distributed dicts
• Environment-level and workspace spend budgets
• Region selection and non-preemptible execution at higher multipliers
• SOC 2 compliance, with HIPAA compatibility on Enterprise

Modal integrations

• AWS Marketplace
• GCP Marketplace
• Slack

How to get started with Modal

1. Sign up and choose a plan; Starter has no subscription fee and includes monthly compute credit.
2. Write a Python function and declare the image, GPU and resources it needs.
3. Deploy it as a web endpoint, scheduled job or batch function, or create a Sandbox for agent code.
4. Modal starts containers on demand, scales them with traffic and shuts them down when idle.
5. Track spend with budgets, tags and billing reports, and invoice monthly.

Modal limitations

• Starter plan includes up to 3 seats, 100 containers and 1 day of log retention
• Custom domains, RBAC and static IP proxy are not on the Starter plan
• Region selection costs 1.15 to 1.75 times base prices
• Non-preemptible execution and Sandboxes are billed at 3 times the base CPU and memory rates
• A payment method must be on file to use Modal

Modal pricing

Yes, Modal has a free plan. Paid plans start at $250 per month, according to the official pricing page on 24 Oct 2026.

Modal plans

• Starter

  $0 + compute /month

  monthly

  • Up to 3 seats
  • $30 / month included compute
  • 1 TiB / month included network egress
  • 100 containers
  • 10 GPU concurrency
  • 1 day log retention

• Team

  $250 + compute /month

  monthly

  • Unlimited seats
  • $100 / month included compute
  • 10 TiB / month included network egress
  • 5000 containers
  • 50 GPU concurrency
  • Custom domains, RBAC and static IP proxy

• Enterprise

  Custom

  • 100 TiB / month included network egress
  • Audit logs, SSO and HIPAA compatibility
  • Use committed spend on AWS and GCP
  • Private Slack support

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

Modal alternatives

The main Modal alternatives listed on Debuly are Replicate, Baseten, Runpod, Together AI, Fireworks AI and Anyscale. The table compares pricing and platforms; order is A–Z.

1. 

   Anyscale

   Managed platform for running Ray workloads, with pay-as-you-go compute and options to deploy in your own cloud.

   Developer platformsAutomation and agents
   PaidVisit Anyscale (official site)
2. 

   Baseten

   Inference platform for deploying open-source, custom and fine-tuned AI models, with per-token Model APIs and dedicated GPUs.

   Developer platforms
   PaidVisit Baseten (official site)
3. 

   Fireworks AI

   Inference and training platform for open models: pay-per-token serverless APIs, GPU deployments, fine-tuning and reserved capacity.

   Developer platformsCodingAutomation and agents
   PaidVisit Fireworks AI (official site)
4. 

   Replicate

   Cloud API for running, fine-tuning and deploying open-source and proprietary AI models, billed by compute time or by output.

   Developer platformsImage generationVideo
   PaidVisit Replicate (official site)
5. 

   Runpod

   Pay-per-second GPU cloud for AI: on-demand Pods, Serverless endpoints, multi-node Clusters and ready-made public model APIs.

   Developer platformsCoding
   PaidVisit Runpod (official site)
6. 

   Together AI

   AI cloud for running open-source models through serverless and batch inference, dedicated endpoints, GPU clusters and fine-tuning.

   Developer platformsCoding
   PaidVisit Together AI (official site)

Modal FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from modal.com · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.