Which is better, Morph or Northflank?
It depends on how busy your sandboxes are and whether you branch them. For agents that mostly wait on a model, Runtime has the lowest bill: $0.64 for 1,000 one-minute runs, against $1.11 on Northflank and $1.67 on Morph, because it bills the CPU the code uses. Pick Morph to clone one running machine into hundreds of replicas. Pick Northflank when every CPU stays busy, where it is cheapest, for GPUs billed by the second, or for your own AWS, GCP or Azure account.
At a glance
| Morph | Northflank | Runtime | |
|---|---|---|---|
| Isolation | Full virtual machines | microVMs for CPU workloads, gVisor for GPU workloads | Firecracker microVM, own kernel, every sandbox |
| CPU | Inside the MCU: 1 vCPU per MCU | $0.01667 per allocated vCPU-hour | $0.025 per vCPU-hour of measured CPU, with a small floor |
| Memory | Inside the MCU: 4 GB per MCU | $0.00833 per GB-hour | $0.0075 per reserved GiB-hour |
| 2 vCPU, 4 GB, an hour | $0.10 (2 MCUs) | $0.0667 | $0.03125 waiting, $0.08 fully busy |
| Plan fee | Developer free with no credit; Team $40 and Scale $250 a month | None published for compute | None; prepaid credit from $10 |
| Free start | No free credit; 1,000 MCUs come with Team ($40) | 2 services, 2 jobs and 1 database free | 100 free hours of a 1 GB machine in the first 30 days, no card |
| Where it runs | Morph Cloud | Northflank's regions, or your AWS, GCP or Azure account | Dedicated servers Runtime operates |
Why does the same machine cost more on Morph?
Northflank's two rates are Morph's MCU split into parts. One MCU, $0.05 an hour, buys 1 vCPU or 4 GB of memory (or 16 GB of disk). A machine pays for as many MCUs as its largest need. On Northflank, 1 vCPU costs $0.01667 an hour, and 4 GB costs $0.03332. Together they cost $0.05.
The two cost the same only when a machine has exactly 4 GB per vCPU. Any other shape pays Morph for resources it does not use:
| Shape | Morph | Northflank |
|---|---|---|
| 2 vCPU, 4 GB | 2 MCUs, $0.10 | $0.0667 |
| 2 vCPU, 8 GB | 2 MCUs, $0.10 | $0.1000 |
| 1 vCPU, 8 GB | 2 MCUs, $0.10 | $0.0833 |
| 4 vCPU, 4 GB | 4 MCUs, $0.20 | $0.1000 |
The MCU does include disk, 16 GB per MCU, which the Northflank figures leave out. Runtime prices CPU and memory apart too. It bills the CPU the code actually uses, so the same 2 vCPU, 4 GiB sandbox costs $0.03125 an hour while it waits on a model.
Cost for the same job
The job: 1,000 runs, 60 seconds each, 20 CPU-seconds of work apiece, on 2 vCPU and 4 GiB. That is 2 MCUs on Morph and 4 GB on Northflank.
- Size
- 1,000 × 60 s / 3,600 × 2 MCU × $0.05
- $1.67
- CPU
- 1,000 × 60 s / 3,600 × 2 × $0.01667
- $0.556
- Memory
- 1,000 × 60 s / 3,600 × 4 × $0.00833
- $0.555
- CPU
- 1,000 × 20 s / 3,600 × $0.025
- $0.14
- Memory
- 1,000 × 60 s / 3,600 × 4 × $0.0075
- $0.50
Across a month of 100,000 runs: $166.67 on Morph, $111.10 on Northflank, and $63.89 on Runtime. With both CPUs busy for the whole minute, the cost is $1.11 on Northflank, $1.33 on Runtime, and $1.67 on Morph. So for saturated CPUs, Northflank's allocated pricing is the cheapest. Plan fees, included credit, storage, network and taxes are left out.
Which one to pick
- Morph, for branching: Infinibranch clones a running machine, memory and all, into hundreds of replicas. Its Devboxes are workspaces you can share with people.
- Northflank, for long CPU-bound jobs, L4, A100 and H100 GPUs billed by the second, or sandboxes inside your own AWS, GCP or Azure account next to other services.
- Runtime, for agents that wait on a model: 62% below Morph and 42% below Northflank on the typical run. It gives a Firecracker microVM per sandbox, forks with memory, and one create call with no service to define.
Moving a workspace
No compatibility import exists for either. The port follows map the calls. A workspace that should last, the job a Morph Devbox or a Northflank service does, becomes a named sandbox. The same call returns it later, waking it if it paused. It pauses on its own when idle:
TypeScriptimport { Sandbox } from "withruntime";const dev = await Sandbox.getOrCreate("dev", { idlePauseSeconds: 900 });await dev.exec("git pull || true");// The next process, or tomorrow, gets the same sandbox and its files.Pythonfrom withruntime import Sandboxdev = Sandbox.get_or_create("dev", idle_pause_seconds=900)dev.exec("git pull || true")npx withruntime compare --from morph or --from northflank reprices your
past usage at Runtime's rates. A coding agent can port the project from the
instruction in migration, on a branch, tested on the
free usage. See Runtime vs Morph,
Runtime vs Northflank and, for workspaces,
per-user dev environments.
Runtime Cloud
Runtime charges 62% less than Morph and 42% less than Northflank when the agent spends its minute waiting. 1,000 one-minute runs of a 2 vCPU, 4 GiB sandbox cost $1.67 on Morph, $1.11 on Northflank, and $0.64 on Runtime. Morph bills the machine's size, and Northflank bills its allocated resources. Runtime bills the CPU the code spends. Rates were checked 25 October 2026.
Run npx withruntime login to start on the free usage. Moving a Morph or
Northflank project? The migration guide maps the calls.
Sources
Rates were checked 25 October 2026. Morph's and Northflank's pricing pages and Northflank's sandbox page were reread 26 October 2026.
- Morph Cloud pricing, Morph Devboxes
- Northflank pricing, Northflank sandboxes
- Runtime pricing and JavaScript SDK
Facts on this page were checked on 26 October 2026.