Modal sandbox

Per second. Billed on max(requested, actual) CPU and memory. Min 0.125 cores / 128 MiB per container. 1 physical core = 2 vCPU.

Cost / code-agent task

$0.0198

2 vCPU / 4 GiB sandbox for 5 minutes at 10% CPU utilization (most time spent waiting on the LLM). Mirrors Vercel's own "AI code validation" example.

1 physical core(s); billed on requested resources

Free tier

Starter: $0/mo with $30/mo included compute.

source ↗ · 2026-10-17

Isolation

gVisor (or microVMs)

source ↗

Rates

Item	Rate	Source	Fetched
Sandbox CPU (physical core = 2 vCPU)	$0.0000394 /core-second	modal.com ↗	2026-10-17
Sandbox memory	$0.00000667 /GiB-second	modal.com ↗	2026-10-17
Region selection	1.15–1.75× multiplier on base price	modal.com ↗	2026-10-17
Non-preemptible	3× multiplier on base price	modal.com ↗	2026-10-17

Plans

Plan	Monthly fee	Included	Concurrency	Max session	Source
Starter	$0	$30/mo compute	100 containers	not listed	source ↗
Team	$250	$100/mo compute	5,000 containers	not listed	source ↗

Limits & details

Seats on Starter: Up to 3 ↗
SDKs: Python; JS/TS and Go (beta) ↗
Pricing page: https://modal.com/pricing
Terms: https://modal.com/legal/terms
Last verified: 2026-10-17

Cold start

No Coldtally benchmark runs published yet. For broad speed coverage see ComputeSDK's open benchmarks (MIT-licensed data). Our methodology explains how we use them.

Changes