Home / Directory / AI infrastructure / Inference & model hosting / Modal
Modal
A serverless compute platform for AI workloads, including elastic inference, agent sandboxes, batch jobs, and training, billed for the time code runs.
Teams that want elastic GPU jobs and agent sandboxes without reserving a cluster.
Buyers that want a long-term dedicated GPU campus lease or a single-model API subscription.
Co-founded by Erik Bernhardsson.
Verdict
Modal is a great fit for developers who need elastic GPU compute for inference, agent sandboxes, batch jobs, or short-lived training. Usage-based pricing is honest: you pay for the time the code runs, and it tends to look good on bursty workloads. Builders like the developer experience and cold-start story; it stands as a sharp elastic compute option, not a multi-year campus lease.
Score Breakdown
How Modal scores in the categories that matter to its buyers.
Pricing
Modal bills GPU, CPU, and memory by the second. Hourly figures below multiply Modal’s published per-second GPU rates by 3,600 and are USD from Modal’s public pricing page as of September 2026. Workspace plans add a monthly fee plus included compute credit. Region and non-preemptible multipliers can raise the effective rate; reserved enterprise deals differ.
| Plan / SKU | Meter | Price (USD) | What stands out |
|---|---|---|---|
| Starter | Workspace | $0 / mo + compute | $30 / mo free compute; 3 seats; 10 GPU concurrency |
| Team | Workspace | $250 / mo + compute | $100 / mo free compute; 50 GPU concurrency |
| Enterprise | Workspace | Sales quote | Higher concurrency, SSO, HIPAA options |
| Nvidia T4 | GPU-second | $0.59 / hr equiv. | $0.000164 / sec |
| Nvidia L4 | GPU-second | $0.80 / hr equiv. | $0.000222 / sec |
| Nvidia A10 | GPU-second | $1.10 / hr equiv. | $0.000306 / sec |
| Nvidia L40S | GPU-second | $1.95 / hr equiv. | $0.000542 / sec |
| Nvidia A100 40GB | GPU-second | $2.10 / hr equiv. | $0.000583 / sec |
| Nvidia A100 80GB | GPU-second | $2.50 / hr equiv. | $0.000694 / sec |
| Nvidia RTX PRO 6000 | GPU-second | $3.03 / hr equiv. | $0.000842 / sec |
| Nvidia H100 SXM5 | GPU-second | $3.95 / hr equiv. | $0.001097 / sec |
| Nvidia H200 SXM | GPU-second | $4.54 / hr equiv. | $0.001261 / sec |
| Nvidia B200 | GPU-second | $6.25 / hr equiv. | $0.001736 / sec |
| Nvidia B300 | GPU-second | $7.10 / hr equiv. | $0.001972 / sec |
The Field at a Glance
How Modal compares with other Inference & model hosting vendors we reviewed, by Overall Score and relative typical engagement cost.
Modal scores 8.4 in Inference & model hosting, ahead of Replicate (8.0), Beam (7.8), Cerebrium (7.8), and Together AI (7.3). Relative cost lands upper-mid for this subcategory.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| Elastic inference | Strong | A published primitive. |
| Agent sandboxes | Strong | Secure environments for code execution. |
| Batch GPU jobs | Strong | Sold beside inference. |
| Long-term dedicated campus lease | Weak | A facilities buy. |
Review Excerpts
Below are excerpts from public reviews. Paid reviews and pay-for-play sites such as Clutch were excluded.
“I like how Modal Labs completely eliminates the infrastructure tax since we don't have to manage Kubernetes, configure CUDA drivers, or manually build Docker containers just to run a Python script on a GPU.”
“Solid offering, a lot of unexpected payments that lead us to switch to another provider”
“Love Modal. We use it for data processing, queues, apis, and all sorts of random things. Such a great product!”
“I've been transitioning some compute-heavy workloads from Lambda/AWS Batch to modal recently and have nothing but good things to say about it. One of those technologies where you are shipping the same afternoon as checking it out. "Wow, that just works?" Highly highly recommended, feels like the future IMO.”
“Modal's usability is amazing, I'm just a bit wary because they use AWS but somehow their prices are lower than that of AWS's GPU machines.”
Company Facts
- Founded
- 2021 (CB Insights)
- Headquarters
- New York, New York, United States (CB Insights)
- Key investors
- Redpoint Ventures, Accel, Lux Capital, Amplify Partners, Menlo Ventures (CB Insights)
- Website
- modal.com
Key Leadership
Erik Bernhardsson
Co-founder and CEO
Methodology
This page is an independent evaluation of Modal for buyers comparing options in inference & model hosting. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. Modal did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria fall under it in two groups.
- Buyer outcomes (for inference & model hosting)
- Elastic inference
- Agent sandboxes
- Batch and training jobs
- Usage-based fit
- Company & commercial
- Innovation & product leadership
- Project management & communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites such as Clutch are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the fit pattern described above.
Scope
Modal is graded here as inference & model hosting. Criteria scores can move as more review volume and product checks are added.