# Runloop vs Prime Sandboxes: sandboxes for evals and training

Runloop and Prime both court agent evaluation and training teams. Runloop is the dearest in the benchmark at $5.33; Prime costs $1.52.

**Runtime runs the benchmark for 88% less than Runloop and 58% less than
Prime.** 1,000 one-minute runs of a 2 vCPU, 4 GiB sandbox, 20 CPU-seconds of
work each: $5.33 on a Runloop `MEDIUM` devbox, $1.52 on Prime Sandboxes, $0.64
on Runtime.

## At a glance

|           | Runloop                                  | Prime Sandboxes                                | Runtime                                                  |
| --------- | ---------------------------------------- | ---------------------------------------------- | -------------------------------------------------------- |
| Isolation | microVM, with a container inside         | Hardware-virtualized microVM, own guest kernel | Firecracker microVM, own kernel                          |
| CPU       | $0.108 per CPU-hour while running        | $0.02 per vCPU-hour while running              | $0.025 per vCPU-hour of measured CPU, with a small floor |
| Memory    | $0.0252 per GB-hour                      | $0.0125 per GiB-hour                           | $0.0075 per reserved GiB-hour                            |
| Evals     | SWE-Bench and custom benchmarks built in | Environments for Prime's RL training           | Harbor and Inspect integrations                          |
| Plan      | Basic free; Pro $250 a month plus usage  | None; rates through 22 December 2026           | None; prepaid credit from $10                            |

## Tooling against price

Runloop sells the benchmark harness: public suites ready to run and custom
ones built from your data, with compliance statements for enterprises. Prime
sells the environment at a low price and huge parallelism. Runtime plugs into
the harnesses teams already use, Harbor with Terminal-Bench and Inspect, and
charges less per rollout than either.

## Cost for the same job

The benchmark, 60 seconds a run (a `MEDIUM` devbox with 8 GB of disk; Prime
with 5 GiB):

```
Runloop  CPU    1,000 × 60 s / 3,600 × 2 × $0.108             = $3.60
         Memory 1,000 × 60 s / 3,600 × 4 × $0.0252            = $1.68
         Disk   1,000 × 60 s / 3,600 × 8 × $0.00034236        = $0.05
         Total                                                  $5.33

Prime    CPU    1,000 × 60 s / 3,600 × 2 × $0.02              = $0.67
         Memory 1,000 × 60 s / 3,600 × 4 × $0.0125            = $0.83
         Disk   1,000 × 60 s / 3,600 × 5 × $0.0002            = $0.02
         Total                                                  $1.52

Runtime  CPU    1,000 × 20 s / 3,600 × $0.025                = $0.14
         Memory 1,000 × 60 s / 3,600 × 4 × $0.0075           = $0.50
         Total                                                  $0.64
```

With both CPUs busy for the whole minute, the job costs $5.33 on Runloop, $1.52 on Prime and $1.33 on Runtime.

## Which one to pick

- **Runloop** for managed benchmark suites and compliance statements.
- **Prime Sandboxes** for RL rollouts by the thousand.
- **Runtime** for eval runs at the lowest price, through
  [Harbor](/docs/harbor) or [Inspect](/docs/inspect).

See [Runtime vs Runloop](/docs/runloop-alternative) and
[Runtime vs Prime Sandboxes](/docs/prime-sandboxes-alternative).

## Sources

Runloop rates checked 23 September 2026; Prime Sandboxes rates checked 25 September 2026.

- [Runloop pricing](https://www.runloop.ai/pricing), [Devbox sizes](https://docs.runloop.ai/docs/devboxes/configuration/sizes)
- [Prime Sandboxes overview](https://docs.primeintellect.ai/sandboxes/overview), [Prime Sandboxes launch](https://www.primeintellect.ai/blog/sandboxes)
- Runtime [pricing](/docs/pricing)

Facts on this page were checked on 25 September 2026.