E2B vs Morph: pricing, isolation and limits

E2B bills vCPUs and memory separately; Morph bills a machine in MCUs, $0.05 each an hour. The same job costs $2.76 on E2B, $1.67 on Morph.

Runtime runs the same job for 77% less than E2B and 62% less than Morph. 1,000 one-minute runs of a 2 vCPU, 4 GiB sandbox cost $2.76 on E2B, $1.67 on Morph and $0.64 on Runtime. Runtime bills the CPU your code uses rather than the machine it holds, and needs no monthly plan. Rates checked 23 September 2026.

At a glance

	E2B	Morph Cloud	Runtime
Isolation	Firecracker microVM, own kernel	Full virtual machines	Firecracker microVM, own kernel
Price	$0.0504 per vCPU-hour, $0.0162 per GiB-hour	$0.05 per MCU-hour: 1 vCPU, 4 GB of memory or 16 GB of disk	$0.025 per vCPU-hour of measured CPU, $0.0075 per GiB-hour
How size is billed	vCPUs plus memory	Whichever of CPU, memory or disk needs the most MCUs	Measured CPU plus reserved memory
Largest machine	8 vCPUs and 8 GiB	64 vCPUs and 256 GB on Developer; 256 vCPUs on Team	vCPUs and memory chosen per sandbox
Plan fee	Hobby $0; Pro $150 a month	Free with no credit; Developer $40 and Team $250 a month	None; prepaid credit from $10
Free start	$100 of usage credit, once	1,000 MCUs with the $40 Developer plan	50 sandbox hours, no card
At once	20 on Hobby, 100 on Pro	8 devboxes on Developer, 32 on Team	8 on the trial, 100 paid to start, raised on request
Branching	Pause keeps files, memory and processes	Infinibranch: a running machine to hundreds of replicas	Forks: 1 to 10 running copies with memory and processes

Is Morph cheaper than E2B?

For a 2 vCPU, 4 GB machine, yes. One MCU covers 1 vCPU, 4 GB of memory or 16 GB of disk, and a machine pays for whichever resource needs the most MCUs. 2 vCPUs need 2 MCUs and 4 GB needs 1, so the machine is 2 MCUs: $0.10 an hour. E2B charges $0.1008 for the two vCPUs and $0.0648 for the memory, $0.1656 an hour.

The MCU rule favours shapes with about 4 GB per vCPU. A 2 vCPU machine with 8 GB is still 2 MCUs on Morph, $0.10 an hour, while E2B adds $0.0648 an hour for the extra memory. A machine with more CPU than memory pays for each vCPU on both.

Morph's plans also count. Its free tier includes no MCUs; the $40 Developer plan includes 1,000 and runs 8 devboxes at once. E2B's free Hobby plan runs 20 and comes with $100 of credit.

Runtime charges for neither the machine nor its allocation: it measures the CPU a run uses, so an agent waiting on a model pays a floor of a twentieth of a vCPU while it waits.

Cost for the same job

1,000 runs of a 2 vCPU, 4 GiB sandbox (2 MCUs on Morph). Each run lasts 60 seconds and keeps the CPU busy for 20 CPU-seconds: an agent that mostly waits for a model.

TextE2B      CPU    1,000 × 60 s / 3,600 × 2 × $0.0504   = $1.68         Memory 1,000 × 60 s / 3,600 × 4 × $0.0162   = $1.08         Total                                          $2.76Morph    Size   1,000 × 60 s / 3,600 × 2 MCU × $0.05 = $1.67Runtime  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

At 100,000 runs a month that is $276.00 on E2B, $166.67 on Morph and $63.89 on Runtime. With both CPUs busy for the whole minute, it is $1.33 on Runtime against $2.76 on E2B and $1.67 on Morph, so Runtime is cheaper however busy the sandbox is. Plan fees, included credit, storage, network and taxes are left out.

Can I branch a running sandbox?

On Morph, yes: Infinibranch copies a running machine, memory included, to hundreds of parallel replicas. E2B pauses and resumes a sandbox with its memory and processes, and keeps a paused one with no expiry.

Runtime forks a running sandbox into 1 to 10 copies, each with the source's files, memory and running processes, and keeps a snapshot to start more later:

Pythonfrom withruntime import Runtimeruntime = Runtime()image = runtime.images.build(name="data", recipe={"pip": ["pandas"], "apt": ["jq"]},                             on_log=lambda line: print(line["text"]))with runtime.sandboxes.create(image=image["id"]) as base:    forks = base.fork(count=2, funding="trial")  # copies as it is now, running    for fork in forks:        fork.stop()    snapshot = base.snapshot(name="with-pandas", retention_days=7)    with runtime.sandboxes.create(snapshot=snapshot["id"]):        pass    runtime.snapshots.delete(snapshot["id"])

Which one to pick

• Pick E2B for vCPUs and memory set separately, a free plan with $100 of credit and 20 sandboxes at once, or a runtime you can self-host under Apache-2.0.
• Pick Morph for machines up to 256 vCPUs, branching a running machine to hundreds of replicas, or shareable Devboxes that scale to zero and keep their memory.
• Pick Runtime for under a quarter of E2B's price and well under half of Morph's on an agent that waits, with no monthly plan, 100 sandboxes at once on a paid account, and forks that keep memory.

Switch from either one

From E2B, Runtime's SDK runs your code after one import change:

TypeScriptimport { Sandbox } from "withruntime/e2b"; // was: from "e2b"

Pythonfrom withruntime.e2b import Sandbox  # was: from e2b import Sandbox

From Morph, a started instance becomes Sandbox.create(), commands run with exec, and stopping becomes stop(). A snapshot you start instances from becomes a Runtime snapshot, and a branch becomes fork:

TypeScriptimport { Sandbox } from "withruntime";const box = await Sandbox.create({ funding: "trial" });try {  console.log((await box.exec("python3 -c 'print(6 * 7)'", { check: true })).stdout);} finally {  await box.stop();}

Or give your coding agent the one instruction in migration: it switches the project on a branch, tests it on the free trial and reports the monthly saving. The full side-by-sides are Runtime vs E2B and Runtime vs Morph. For evals that fan out from one prepared machine, see agent evals and SWE-bench.

Sources

E2B and Morph pricing pages rechecked 25 September 2026; the rest checked 23 September 2026.

• E2B pricing, E2B sandbox persistence
• Morph Cloud pricing, Morph Devboxes, Morph snapshots
• Runtime pricing and Python guide

Facts on this page were checked on 25 September 2026.