Side by side
| Modal Sandboxes | Uno | |
|---|---|---|
| What it is | Sandboxes you create from the Modal SDK, next to its functions and GPUs | A Linux computer with root and Docker, plus sites, storage and Uno AI |
| Isolation | gVisor by default, or a VM runtime for work that needs a full Linux, such as Docker inside the sandbox | Firecracker microVM, one per computer, root inside |
| How you pay | Starter: $0 a month plus compute, with $30 of credits every month. Team: $250 | A flat plan: Small $5, Plus $20 |
| Rates | $0.141912 per core-hour and $0.024012 per GiB-hour. Modal bills per physical core, and one core is 2 vCPU equivalent | None on top of the plan |
| 1 vCPU · 1 GiB, a month | $69.33 | $5 |
| 4 vCPU · 4 GiB, a month | $277.31 | $20 |
| Longest run | 5 minutes by default, up to 24 hours | No session time limit |
| When idle | A sandbox ends by itself after a stretch without activity when you set idle_timeout; running a command, writing to stdin or keeping a TCP tunnel open counts as activity | Asleep after 10 minutes; memory and running programs are saved while it sleeps |
| Keeping state | Filesystem snapshots stay 30 days; memory snapshots 7 days, and memory snapshots are experimental, and Modal lists known limitations | Disk kept, and memory and running programs are saved while it sleeps |
| What wakes it | Your code starts a new sandbox, from a snapshot if you took one | Wakes when you open it, a visitor opens your site, a message reaches your bot's webhook |
| Smallest size | 0.125 cores | 1 vCPU and 1 GB on Small |
| GPUs | The sandbox docs list GPU allocation next to CPU and memory | None; plans list vCPU and memory |
One machine busy for a month, Modal against Uno
Whole-month cost in $ per month: 730 hours at the Modal rates, against the flat Uno plan. A Modal core is 2 vCPU, so the first Modal bar is half a core and the second a whole one.
Show as a table
| Option | $ per month | Shape |
|---|---|---|
| Uno Small, 1 vCPU · 1 GB | 5 | Flat price, sleeps when idle, keeps memory |
| Modal, 1 vCPU · 1 GiB | 69.33 | Half a core plus 1 GiB, 730 hours |
| Modal, 1 core (2 vCPU) · 1 GiB | 121.12 | One whole core plus 1 GiB, 730 hours; a run still ends within 24 hours |
| Uno Plus, 4 vCPU · 4 GB | 20 | Flat price, sleeps when idle, keeps memory |
| Modal, 4 vCPU · 4 GiB | 277.31 | Two cores plus 4 GiB, 730 hours |
How we counted
730 hours in a month, all busy. Half a core and 1 GiB: 730 × (0.5 × $0.141912 + $0.024012) = $69.33. A whole core and 1 GiB: 730 × ($0.141912 + $0.024012) = $121.12. Two cores and 4 GiB give $277.31. Modal cannot hold one sandbox for the whole month: each ends within 24 hours, so a month means about thirty sandboxes in a row, with the state carried by snapshots.
The clock on a sandbox
A Modal sandbox has a lifetime. It starts at 5 minutes and you can raise it to 24 hours. After that the sandbox is gone, and what you want to keep must already be in a snapshot. Taking a snapshot ends the sandbox, and open TCP connections close and have to be opened again. An Uno computer has no such clock: it sleeps when idle and wakes as the same computer.
For a job that ends in an hour, none of this matters, and per-second billing is a good deal. For an agent that holds a project open for a week, it means a handoff every day.
Where pay-per-use wins
An hour of 1 vCPU (half a Modal core) and 1 GiB of memory costs $0.0950 on Modal, so $5 buys about 53 busy hours. Under that, Modal is cheaper; over it, the flat plan is. Plus lands on the same number, since its price and its size are both four times Small's.
Where Uno is worse
- Under about 53 busy hours a month, Modal costs less than $5, and Starter has no monthly fee.
- Take Modal for fleets: it is built to create many sandboxes from code, with GPUs available next to them. An Uno computer is meant for one agent that stays.
- Modal's default runtime is gVisor, and it offers a VM runtime when you need Docker inside. Uno is a microVM with root and Docker in every plan.
- Modal's smallest size is 0.125 cores. Our smallest computer is 1 vCPU and 1 GB.
- Our vCPU are shared, not dedicated, and paid plans run in one region today: Netherlands.
Your agent gets a clean address of its own
Your computer runs in the Netherlands. Its traffic can leave through a static address of a regular internet provider in one of 8 countries. Only your account works from that address.
- From a data-center address
- Sites see a server range shared with thousands of strangers. The agent's browser hits more captchas and blocks.
- From your Clean exit IP
- Sites see a regular internet provider's address in the country you picked. Fewer captchas, and less reason for Claude or ChatGPT to flag your account: nobody else works from it.
A computer that keeps going, from $5 a month
Every plan is a Linux microVM with root and Docker. Pick the size; move up or down any time.
Small
A Telegram bot, a VPN, a small backend
$5/ month
- 1 vCPU · 1 GB
- 15 GB disk
- 25 GB cloud storage
- 5 boosts a month
Plus
Agents that work with your laptop closed
$20/ month
- 4 vCPU · 4 GB
- 40 GB disk
- 100 GB cloud storage
- Uno Work in the cloud
Pro
A product in production, a second MacBook
$70/ month
- 16 vCPU · 16 GB
- 120 GB disk
- 300 GB cloud storage
- Speed of a MacBook Air M4
Max
A team and many agents at once
$200/ month
- 32 vCPU · 48 GB
- 300 GB disk
- 1,024 GB cloud storage
- 5 public IPv4
Add Uno AI to any plan: 40 h of AI work a month on Small, 100 h on Plus, unlimited on Max. Today: GLM-5.3 Flash and MiMo v2.6 Pro. Or sign in with your own Claude or ChatGPT.