Beam
Choose Beam after deciding whether the AI workload needs a persistent machine, a short sandbox or a serverless GPU endpoint.
Overview
What you provide
- Provide your application code or machine configuration, dependencies and the CPU, RAM or GPU requirement.
What you get
- Beam runs the selected compute workload and exposes its application results through the configured service or machine.
- Starting price
- Base VM from USD 0.0306/hour
- Free access
- No quantified recurring free allowance verified
- Billing unit
- Resource runtime; different rates by workload
- Platform
- Hosted Linux VMs, serverless GPUs and SDK
- Verified
- 4 Nov 2026
Persistent VMs start at USD 0.0306/hour for one vCPU, 1 GiB RAM and an included 50 GiB root disk, billed per second while running. Serverless RTX 4090 GPU time is USD 0.000192/second, with GPU-attached CPU USD 0.000105/physical core/second and RAM USD 0.0000055/GiB/second. Sandboxes use USD 0.0000375/core/second plus USD 0.0000064/GiB/second. Persistent volumes include up to 1 TB; additional storage is USD 0.021/GB/month. No quantified general free credit was verified in this capture.
Official pricing ↗Pricing checked:
Setup & workflow
- Identify the required lifetime and choose VM, sandbox or serverless execution.
- Test a representative job and measure its active resource time.
- Review compute and persistent storage separately before scaling.
Requirements & installation
- Prepare a deployable workload and select a resource configuration rather than assuming a hosted model is included.
Limits & review
- On-demand machine prices and serverless component prices are different products. Persistent data can outlive stopped compute, so stopping a VM does not establish a zero total bill.
Your part
- Measure cold starts and actual resource use, and verify model outputs before relying on the deployed endpoint.
Frequently asked questions
What is Beam?
Beam is a compute service for AI workloads using persistent machines, sandboxes or serverless GPU endpoints.
What can Beam do?
Beam runs the selected compute workload and exposes its application results through the configured service or machine.
Who is Beam for?
It is for people selecting persistent machines, short sandboxes or serverless GPU endpoints for AI workloads.
How much does Beam cost?
Persistent VMs start at USD 0.0306/hour for one vCPU, 1 GiB RAM and an included 50 GiB root disk, billed per second while running; Serverless RTX 4090 GPU time is USD 0.000192/second, with GPU-attached CPU USD 0.000105/physical core/second and RAM USD 0.0000055/GiB/second; Sandboxes use USD 0.0000375/core/second plus USD 0.0000064/GiB/second. Persistent volumes include up to 1 TB; additional storage is USD 0.021/GB/month. No quantified general free credit was verified in this capture.
What should I check before using Beam?
Before starting, check these setup requirements: Prepare a deployable workload and select a resource configuration rather than assuming a hosted model is included.
When should I look at something else?
Consider alternatives if this documented limit conflicts with your needs: On-demand machine prices and serverless component prices are different products. Persistent data can outlive stopped compute, so stopping a VM does not establish a zero total bill.
Choosing this tool
The low VM headline is not a GPU endpoint tariff. Beam publishes distinct CPU and memory rates for CPU-only work, GPU-attached work and sandboxes, so keep those configurations separate in a budget. Cerebrium is relevant for a per-second deployed AI service with a platform tier; Daytona is relevant when the main requirement is isolated agent execution. Choose using duty cycle and resource requirements rather than a single hourly number.
A decision worked through
The base one-vCPU VM running 160 hours costs 160 × USD 0.0306 = USD 4.896 before separately charged services. A serverless GPU endpoint must instead add GPU, attached CPU and RAM costs; this VM example cannot price it.
Beam alternatives for serverless ai compute and persistent agent machines
Compare Beam with Cerebrium, Baseten for this workflow, including the billing unit and the work left to your team.
| Product | Price / unit | Free allowance | Input → output | Key limit |
|---|---|---|---|---|
| BeamRead profileOfficial source ↗ | VM from USD 0.0306/hour for 1 vCPU and 1 GiB | General free-credit quantity unverified | Code and resources become running AI services or machines. | GPU-attached CPU and RAM are extra components. |
| CerebriumRead profileOfficial source ↗ | Hobby USD 0 platform fee; Standard USD 100/month | First 100 GB storage free; compute billed | Application code becomes hosted inference endpoints. | Public seat and log terms conflict; confirm them. |
| BasetenRead profileOfficial source ↗ | T4 dedicated compute USD 0.01052/minute | Quantified free grant not established | Models and code become hosted inference or execution services. | Model token prices differ from allocated compute rates. |
Website traffic
beam.cloud
All available locations and languages
Estimated monthly paid search traffic: 0
Snapshot captured:
Search traffic is estimated from keyword rankings, search volume and click-through rates. It excludes direct and referral visits and is not total website traffic.
Monthly values
| Month | Estimated visits |
|---|---|
| 2025-10 | — |
| 2025-11 | 2,158 |
| 2025-12 | 2,304 |
| 2026-01 | 4,401 |
| 2026-02 | 4,546 |
| 2026-03 | 4,648 |
| 2026-04 | 5,902 |
| 2026-05 | 5,896 |
| 2026-06 | 1,897 |
| 2026-07 | 1,473 |
| 2026-08 | 2,569 |
| 2026-09 | 2,770 |
| 2026-10 | 2,410 |
Months without an estimate are left blank, not counted as zero.
Domain-wide estimate; includes other products and pages on this domain. This is not verified usage, revenue or product quality.
Sources & verification5
This profile is based on official sources, not a hands-on product test.
Vendor descriptions and demos document advertised features. Editorial guidance is based on these sources.
- Official Beam pricingwww.beam.cloud
- Official Beam productbeam.cloud
- Beam official referencewww.beam.cloud
- Cerebrium official referencewww.cerebrium.ai
- Baseten official referencewww.baseten.co
Product details
- Product type
- AI tool
- Getting started
- Ready to use
- Category
- Coding Agents & App Builders
- Evidence
- Official-source review