Render vs Railway: pricing
Render and Railway both bill by resource-time rather than per request. Render is web services / containers (instance-hour) (Not request-priced) and Railway is usage-based containers (per-minute vcpu + ram) (Not request-priced). Snapshot June 2026 — verify on the vendor's pricing page.
Data as of June 2026.
Render vs Railway side by side
| Cost factor | Render | Railway |
|---|---|---|
| Vendor | Render | Railway |
| Pricing model | Web services / containers (instance-hour) | Usage-based containers (per-minute vCPU + RAM) |
| Per 1M requests | Not request-priced | Not request-priced |
| Per GB-second | n/a | n/a |
| Free tier | Free web service: 750 instance-hours/month (512 MB / 0.1 CPU), spins down after 15 min idle with a 30-60 s cold start. Paid Starter from $7/month (always-on, 512 MB) up to $85/mo Pro (4 GB / 2 CPU). | No perpetual free runtime. Hobby is $5/month including $5 of usage. Usage-based: $20/vCPU/month ($0.000463/vCPU/min) + $10/GB-RAM/month ($0.000231/GB/min), billed per minute. Egress $0.10/GB. |
| Scale to zero | Yes | No |
| Egress | varies | $0.10/GB |
| Regions | Oregon, Ohio, Frankfurt, Singapore. | US, EU and Asia-Pacific regions. |
Sources: Render pricing · Railway pricing.
Sample monthly cost
A rough monthly total for a sample workload: 5,000,000 requests/month, each ~150 ms at 256 MB of memory, plus 25 GB of egress. Computed after each platform's free tier; excludes platform minimums and regional variation, so treat it as a starting point:
| Component | Render | Railway |
|---|---|---|
| Compute (after free tier) | Usage-based | Usage-based |
| Egress (25 GB) | varies | $2.50 |
| Estimated total / month | Usage-based | Usage-based |
Verdict
On cost for this sample workload, Render and Railway both bill by resource-time rather than per request. But price is not the only factor: Render (simple instance pricing, no per-request math) versus Railway (per-minute usage billing, pay for what runs). Read each platform's full breakdown — Render and Railway — and model your own workload in the calculator.
Frequently asked questions
Is Render or Railway cheaper?
For a sample workload of 5M requests/month at 256 MB, 150 ms each plus 25 GB egress, Render and Railway both bill by resource-time rather than per request. This is illustrative and after each platform's free tier; your real cost depends on memory, duration, egress and region. snapshot captured June 2026.
How do Render and Railway price requests?
Render: Instance-based: Free tier gives 750 instance-hours/month and spins down after 15 min idle (cold start 30-60 s). Paid web services from $7/mo (Starter, always-on) to $85/mo (Pro, 4 GB/2 CPU). Not per-request priced. Railway: Usage-based, billed per minute: $20 per vCPU/month ($0.000463/vCPU/min) + $10 per GB-RAM/month ($0.000231/GB/min). Hobby plan $5/mo includes $5 usage. Egress $0.10/GB; volumes ~$0.25/GB-mo. Not per-request priced. Snapshot June 2026 — verify on the vendor's pricing page.
Do Render and Railway scale to zero?
Render: yes, no charge when idle. Railway: no, it bills while running. Scale-to-zero saves money for spiky traffic but adds cold-start latency on the first request.
Which has the better free tier, Render or Railway?
Render: Free web service: 750 instance-hours/month (512 MB / 0.1 CPU), spins down after 15 min idle with a 30-60 s cold start. Paid Starter from $7/month (always-on, 512 MB) up to $85/mo Pro (4 GB / 2 CPU). Railway: No perpetual free runtime. Hobby is $5/month including $5 of usage. Usage-based: $20/vCPU/month ($0.000463/vCPU/min) + $10/GB-RAM/month ($0.000231/GB/min), billed per minute. Egress $0.10/GB.
More comparisons
Last updated: 2026-06-21