Fly Sprites vs Blaxel: pricing, standby and limits

Both pause idle sandboxes by themselves. Sprites bill measured CPU and memory in use; Blaxel bills memory, with CPU included.

Runtime runs the same job for 81% less than Sprites and 77% less than Blaxel. 1,000 one-minute runs of a 2 vCPU, 4 GiB sandbox cost $3.31 on Fly Sprites, $2.76 on Blaxel and $0.64 on Runtime. Runtime also pauses idle sandboxes and wakes them on the next request, and its paused storage costs less than Blaxel's standby snapshots. Rates checked 23 September 2026.

At a glance

	Fly Sprites	Blaxel	Runtime
Isolation	Firecracker microVM	Lightweight virtual machines	Firecracker microVM, own kernel
CPU	$0.07 per CPU-hour of measured CPU	Included with memory; 8 GB gets 4 cores	$0.025 per vCPU-hour of measured CPU, with a small floor
Memory	$0.04375 per GB-hour of memory in use	$0.0000115 per GB-second of active time	$0.0075 per reserved GiB-hour
Goes idle	About 30 seconds after the last activity	About 15 seconds after the last connection closes	After idlePauseSeconds you set, 60 to 86,400
Idle storage	Cold disk $0.000027 per GB-hour, about $0.02 a GB-month	Standby snapshot $0.20 per GB-month	Paused storage $0.08 per GB-month
Plan fee	None required; optional monthly plans	None; tiers grow with credit added	None; prepaid credit from $10
Free start	$30 of trial credit	Up to $200 of credit	50 sandbox hours, no card
At once	Not stated on the pricing page	10 free, 50 at $20 a month of top-ups, 200 at $50	100 sandboxes on a paid account to start

What survives when a sandbox goes idle?

On Blaxel, everything inside the VM. A sandbox goes to standby about 15 seconds after its connections close (open WebSocket and TCP connections time out after 15 idle minutes). Blaxel snapshots the whole state, filesystem and running processes, and says reconnecting resumes it in under 25 ms. Connections the sandbox held open to a database or an API do not survive. Blaxel keeps the writable filesystem in memory, using about half of the sandbox's memory for it.

On Sprites, the disk always, memory for a while. A Sprite pauses about 30 seconds after activity stops. In a warm pause its memory is frozen and it resumes in 100 to 500 ms with processes carrying on. Left longer, it goes cold: memory is dropped, the next wake takes 1 to 2 seconds, and processes start fresh. Fly does not state when a warm pause turns cold. The 100 GB disk is kept, and sleeping storage is cheap.

On Runtime, files, memory and processes, for as long as you choose. A paused sandbox is kept 1 to 365 days on a paid account, at $0.08 per GB a month, and a request to it wakes it in about half a second. Nothing drops its memory on a schedule you did not set.

Cost for the same job

1,000 runs of a 2 vCPU, 4 GiB sandbox (4 GB on both rivals). Each run lasts 60 seconds and keeps the CPU busy for 20 CPU-seconds: an agent that mostly waits for a model. The Sprite is assumed to use all 4 GB for the whole minute.

TextSprites  CPU    1,000 × 20 s / 3,600 × $0.07         = $0.39         Memory 1,000 × 60 s / 3,600 × 4 × $0.04375  = $2.92         Total                                          $3.31Blaxel   Memory 1,000 × 60 s × 4 × $0.0000115        = $2.76Runtime  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 $330.56 on Sprites, $276.00 on Blaxel and $63.89 on Runtime. With both CPUs busy for the whole minute, it is $1.33 on Runtime against $5.25 on Sprites and $2.76 on Blaxel, so Runtime is cheaper however busy the sandbox is. Busy work costs Sprites more because its CPU is metered; Blaxel's price does not move with the CPU. The figures assume each sandbox stops billing the moment its run ends, where Sprites stay awake about 30 seconds and Blaxel about 15. Storage, network, plan fees and free credits are left out.

Which one to pick

• Pick Fly Sprites for a persistent computer an agent keeps coming back to, with a 100 GB disk that costs little while it sleeps and filesystem checkpoints taken in about a second.
• Pick Blaxel if the sandbox must come back in milliseconds with its processes running, or for very large fleets: its top tier runs over 100,000 sandboxes at once.
• Pick Runtime for the lowest cost of the three while running, pause that keeps memory for the retention you choose, and forks of a running sandbox.

Switch from either one

Both rivals pause on their own; Runtime does the same when you set an idle time, and any request wakes the sandbox. Keep one per user or task by name:

TypeScriptimport { Sandbox } from "withruntime";const dev = await Sandbox.getOrCreate("agent-42", { idlePauseSeconds: 60 });await dev.exec("git pull || true"); // wakes it first if it was paused

From Sprites, client.createSprite(name) becomes Sandbox.create() (or Sandbox.getOrCreate(name) to reach it by name again) and sprite.execFile(cmd, args) becomes exec([cmd, ...args]). Blaxel's standby maps to automatic wake, which is on by default. 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. npx withruntime compare --from fly (or --from blaxel) prices what you ran at the old provider's rates, and npx withruntime switch --from fly before your first top-up matches that top-up with credit, up to $100. The full side-by-sides are Runtime vs Fly.io and Runtime vs Blaxel.

Sources

Rates checked 23 September 2026; the Sprites, Sprites lifecycle, Blaxel pricing and Blaxel sandbox pages read 25 September 2026.

• Fly.io pricing, Sprites, Sprites lifecycle
• Blaxel pricing, Blaxel sandboxes
• Runtime pricing and how to pause and resume a sandbox

Facts on this page were checked on 25 September 2026.