Daytona sandbox
Per second, pay-as-you-go.
Cost / code-agent task
$0.0138
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.
Rates
| Item | Rate | Source | Fetched |
|---|---|---|---|
| vCPU | $0.0504 /vCPU-hour | www.daytona.io ↗ | 2026-10-17 |
| Memory | $0.0162 /GiB-hour | www.daytona.io ↗ | 2026-10-17 |
| Storage (after first 5 GiB free) | $0.000108 /GiB-hour | www.daytona.io ↗ | 2026-10-17 |
| Windows OS | $0.0858 /vCPU-hour | www.daytona.io ↗ | 2026-10-17 |
Plans
| Plan | Monthly fee | Included | Concurrency | Max session | Source |
|---|---|---|---|---|---|
| Pay as you go | $0 | $200 free compute (one-time) | not listed | not listed | source ↗ |
| Startups | Custom | Up to $50k credits (application) | not listed | not listed | source ↗ |
Limits & details
- Concurrency / session limits
- Unverified: Not listed on the pricing page; not verified. ↗
- SDKs
- Python, TypeScript, Go, Ruby, Java ↗
- Pricing page
- https://www.daytona.io/pricing
- Terms
- https://www.daytona.io/terms-of-service
- Last verified
- 2026-10-17
Not yet verified
- Not listed on the pricing page; not verified.
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
No changes recorded since the baseline (2026-10-17).
What changed in agent infra, weekly
Every pricing, plan and limit change we detect for sandboxes and cloud browsers, with sources. One email a week. No vendor-paid content.