# Render

Official pricing: https://render.com/pricing  
Category: paas · Isolation: container

## Pricing regimes (raw)

- **Web / private service or background worker (fixed compute plan, prorated per second)** (sizes): 0.5c-512mb (Starter) 0.5 vCPU/0.5 GiB $0.009589/h; 1c-2g (Standard) 1 vCPU/2 GiB $0.034247/h; 2c-4g (Pro) 2 vCPU/4 GiB $0.116438/h; 2c-8g 2 vCPU/8 GiB $0.184932/h; 2c-16g 2 vCPU/16 GiB $0.273973/h; 4c-8g (Pro Plus) 4 vCPU/8 GiB $0.239726/h
- **Cron job (billed per second while running)** (sizes): 0.5c-512mb 0.5 vCPU/0.5 GiB $0.0096/h; 1c-2g 1 vCPU/2 GiB $0.0348/h; 2c-4g 2 vCPU/4 GiB $0.1182/h; 2c-8g 2 vCPU/8 GiB $0.1878/h; 2c-16g 2 vCPU/16 GiB $0.2778/h; 4c-8g 4 vCPU/8 GiB $0.243/h [alt]
- **Workflow task run, fixed compute plan (per second while running)** (sizes): 2c-4g 2 vCPU/4 GiB $0.4/h; 2c-8g 2 vCPU/8 GiB $0.7/h; 4c-8g 4 vCPU/8 GiB $1/h; 4c-16g 4 vCPU/16 GiB $1.5/h; 8c-32g 8 vCPU/32 GiB $2.5/h [alt]
- **Workflow task run, flex plan (billed on actual CPU time + sampled RAM)** (resource): $0.2/vCPU-h, $0.05/GiB-h [alt]
- **Render Sandboxes (early access, price unpublished)** (sizes): sandbox (EA limit 2 CPU / 4 GB / 10 GB) 2 vCPU/4 GiB $null/h [beta, sales]

## Features

Yes: snapshots, fork/clone, pause/resume, persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), start from your own snapshot, your own Docker/OCI image, SSH, HTTPS preview URLs, custom domains, egress allowlist, open internet, static egress IP, private networking, SOC 2, HIPAA, SSO, Python, Node.js, Go, Rust, live fork (no pause), memory fork, webhooks, scheduled runs (cron), audit logs, EU data residency, MCP server, snapshots on demand, inbound access rules, extra volumes

No: memory snapshots, idle auto-stop, full VM (own kernel), Docker inside, nested virtualization, systemd, browser, desktop GUI, computer-use API, code interpreter, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, public IPv4, raw TCP inbound, self-hosting, BYOC, open source, GPU, arm64, Windows, macOS, GPU desktop, HTTP method/path egress rules, wake on request, live resize, spend limits, automatic snapshots, agent harness API, hosted agent API (their own agent), gVisor or VM (no shared kernel), firewall inside (nftables), secret proxy, secret proxy for any API, shared volumes

Unknown: everything else. Evidence (source + quote) per feature: https://battleships.dev/data/providers/render.json → feature_evidence

## Caveats

- Render is a PaaS: 'sessions' map to services you create and suspend yourself. Paid compute never auto-stops; per-second proration only helps if you suspend or delete services.
- hour = monthly price / 730 and month_cap = monthly price are derived; Render only says 'prorated to the second' and does not state the month length used.
- Plans carry a non-spec 'egress_free_gib' key: included bandwidth is per workspace plan (5 GB Hobby / 25 GB Pro / 1 TB Scale); card network block uses the Hobby value. Overage $0.15/GB public, $0.03/GB via private link; inbound, same-region private network and same-region S3/GCS traffic free. Without a card, Render suspends services instead of billing overage.
- Hobby concurrency 25 = max services per workspace, not an instance limit; scaled services bill every instance.
- snapshot_gib_month = persistent-disk price ($0.25/GB-mo): Render has no priced snapshot product; daily disk snapshots (kept >= 7 days, full restore only) have no separate published charge. Disk billed while the service exists (incl. suspended; assumed from FAQ 'storage capacity' billing).
- Persistent disk pins a service to a single instance and makes compute-plan changes require downtime.
- No public IPv4 per instance; dedicated outbound IPs are $100/month per IP set (Pro+).
- Render Sandboxes (early access, 2 CPU/4 GB/10 GB, Oregon only, runtime memory snapshots) has NO published price; mode kept with hour null and flags beta+sales.
- HIPAA: +20% on all usage via knob; Singapore unavailable for HIPAA workspaces.
- Workflows/cron are batch products (flag alt), not interactive machines; workflow flex caps at 1 CPU / 4 GB.
- Compute plans renamed in Aug 2026 (Starter/Standard/Pro/Pro Plus/Pro Max/Pro Ultra -> 0.5c-512mb ... 8c-32g) with no price change; new sizes 2c-16g, 8c-64g, 12c-48g added.
- Startup credits (up to $100K) and migration credits (up to $10K) are application-based and not modelled.
- Postgres, Key Value, build minutes ($5/1K, performance $25/1K) and custom domains ($0.25/mo) not modelled.
- No benchmark data for Render in the research set (perf null). Isolation 'container' per Render docs ('containerized environments').
- Workflows pricing changed at GA on 2026-09-01: workspace-wide run-concurrency limits and +$10/50 add-ons removed, replaced by per-workflow resource limits; flex $0.20/active CPU-h + $0.05/active GB-h with a 1 s / 0.1 CPU / 0.1 GB minimum; task state $0.25/GB-month. Flat Pro $25 / Scale $499 workspace plans were opt-in from 2026-04-23 and applied to all workspaces 2026-08-01. (patch v8-history, re-checked render.com/blog/upcoming-changes-to-workflows-pricing 2026-09-28)
- Free web service tier (0.1 CPU / 512 MB, spins down after 15 min idle, 750 instance-hours/workspace/month, not for production; render.com/docs/free) removed from priced regimes on 2026-09-28: it cannot run any preset's workload and was being clamped into $0-compute rows.
- Per-service scaling cap: 100 instances for compute plans up to 8c-32g, but only 5 instances for 8c-64g, 12c-48g and 12c-96g (contact Render to exceed).
- Preview environments and PR service previews bill as full services, prorated per second (same compute plan as base unless overridden); previews.expireAfterDays default is no expiry.
- render.com/pricing comparison table still shows workflow 'Concurrent task runs 20 / 50 / 100, then $10/mo per 50' and 'Task compute plans Up to Pro / Up to Pro Ultra'; docs/workflows-limits (and the 2026-09-01 changelog) replaced this with per-minute compute limits. Card follows the docs; the pricing page is stale.

## How this provider charges


Render bills three things: a flat **workspace plan** fee, **compute** (fixed plans with a monthly list price,
prorated to the second while an instance runs), and **metered extras** (outbound bandwidth over the plan allowance,
disks, pipeline minutes, domains). Same prices in all 5 regions (Oregon, Ohio, Virginia, Frankfurt, Singapore).
Paid services never auto-stop, so "per second" only helps if you suspend/scale down yourself. The sandbox-like
product (Render Sandboxes) is early access with **no published price**.

All numbers verified live 2026-09-28 on render.com/pricing and render.com/docs (pages saved in scratchpad; no
third-party sources used).

4c-8g (4 CPU / 8 GB) = $175/month ≈ **$0.2397/h** (175/730; Render doesn't state the proration month length).

## Regime table

| Regime | When it applies | How billed | Numbers | Source |
|---|---|---|---|---|
| Hobby workspace | Default | No fee | $0; 1 seat; max 25 services; 5 GB bandwidth; 500 pipeline min; 2 domains; no autoscaling | https://render.com/pricing |
| Pro workspace | Teams / production | Flat per workspace (not per seat) + compute; waived in a month with no services and no activity | $25/mo; unlimited seats & services; 25 GB bandwidth; 1K pipeline min; 15 domains | https://render.com/pricing |
| Scale workspace | Governance / HIPAA / SSO | Flat + compute | $499/mo; 1 TB bandwidth; 5K pipeline min; multiple workspaces; SAML/SCIM | https://render.com/pricing |
| Enterprise | SLAs, TAM | Custom | null | https://render.com/pricing |
| Service compute (web, private, worker) | Any paid instance | Monthly price prorated per second while running; each instance of a scaled service and each preview instance billed; suspended = not billed | 0.5c-512mb $7, 1c-2g $25, 2c-4g $85, 2c-8g $135, 2c-16g $200, 4c-8g $175, 4c-16g $225, 4c-32g $350, 8c-16g $300, 8c-32g $450, 8c-64g $1,000, 12c-24g $450, 12c-48g $800, 12c-96g $1,500 per month | https://render.com/pricing, https://render.com/docs/compute-plans, https://render.com/docs/faq |
| Always-on service | Instance runs all month | Capped at the monthly list price (hourly proration × 730) | = list prices above | https://render.com/pricing |
| Free instance | Web services (and free Postgres/Key Value) | $0; 750 instance-hours/workspace/month; spins down after 15 min without inbound traffic (~1 min spin-up); all free web services suspended when hours run out | 0.1 CPU / 512 MB; no disk; free Postgres expires after 30 days | https://render.com/docs/free |
| Cron job | Scheduled command | Per-minute price, prorated per second while running | $0.00016/min (0.5c) … $0.00405/min (4c-8g = $0.243/h) … $0.02315/min (8c-64g) | https://render.com/pricing |
| Workflows, fixed plans | Durable task runs | Per second while running; default 2 h timeout, up to 24 h | 2c-4g $0.40/h, 2c-8g $0.70/h, 4c-8g $1.00/h, 4c-16g $1.50/h, 8c-32g $2.50/h | https://render.com/docs/workflows-limits |
| Workflows, flex (default) | Task runs ≤ 1 CPU / 4 GB | Actual CPU time + sampled RAM | $0.20/CPU-h + $0.05/GB-h (max $0.40/h); min charge = 1 s at 0.1 CPU/0.1 GB | https://render.com/docs/workflows-limits |
| Workflow concurrency | Concurrent task runs | Included per plan + add-on | 20 Hobby / 50 Pro / 100 Scale; +$10/mo per 50 | https://render.com/pricing |
| Workflow task state | Inputs/return values retained 30 days | Billed once per run | $0.25/GB | https://render.com/docs/workflows-limits |
| Render Sandboxes | Early access, select workspaces by request | **Unpublished** | EA limits: 2 CPU / 4 GB / 10 GB per sandbox, 100 concurrent, 24 h timeout, Oregon only; filesystem or runtime (memory) snapshots, expire after 3 days | https://render.com/docs/sandboxes |
| Persistent disk | Attached SSD | Per GB-month of provisioned size; daily snapshots (≥ 7 days) no separate charge published | $0.25/GB-mo | https://render.com/pricing, https://render.com/docs/disks |
| Outbound bandwidth | Traffic to outside Render | Over plan allowance; unused doesn't roll over; no card → services suspended | $0.15/GB public; $0.03/GB via private link; inbound/private-network/same-region S3-GCS free | https://render.com/docs/outbound-bandwidth |
| Dedicated outbound IPs | Pro+ | Monthly | $100/mo per IP set | https://render.com/pricing |
| AWS PrivateLink | Pro+ | Monthly | $30/mo for up to 3 links + $0.03/GB | https://render.com/pricing |
| Build pipeline | Builds + pre-deploy | Over plan allowance | $5/1K min standard; $25/1K min performance (no allowance); spend limit available | https://render.com/pricing, https://render.com/docs/faq |
| HIPAA-enabled workspace | Scale+ with BAA; irreversible | +20% on all usage (compute, storage, …); no Singapore | ×1.2 | https://render.com/docs/hipaa-compliance |
| Credits | Startups / migrations | Application-based | up to $100K (startups), up to $10K (migration) | https://render.com/pricing |

Dated change: Aug 2026 compute plans renamed (Starter → 0.5c-512mb, Standard → 1c-2g, Pro → 2c-4g, Pro Plus → 4c-8g,
Pro Max → 4c-16g, Pro Ultra → 8c-32g) with "no pricing changes to existing plans"; 2c-16g, 8c-64g, 12c-24g/48g/96g
added. Workflows beta starter/standard plans replaced by flex.

## Gotchas

1. **No idle auto-stop on paid compute.** Only Free web services spin down. An agent box left running bills up to the
   full monthly price; stop the meter with suspend (compute not billed while suspended) or scale to 0.
2. **Disks pin you to one instance** and force downtime on plan changes; disk GB is billed while the service exists.
3. **Plan fee is not credit** and is per workspace, not per seat (Pro $25 with unlimited seats).
4. **Bandwidth is per workspace plan** (5 GB / 25 GB / 1 TB), then $0.15/GB. Without a card on file, Render suspends
   services instead of charging.
5. **Previews and every scaled instance bill** at the full plan rate.
6. **Workflows cost ~4x services per hour** (4c-8g $1.00/h vs $0.24/h) but scale to zero; flex is cheap only for
   ≤ 1 CPU work.
7. **HIPAA is +20% on everything** and irreversible; Singapore unavailable.
8. **Sandboxes have no price** and are Oregon-only early access — can't be used for cost planning yet.
9. **Free tier is time-boxed**: 750 instance-hours per workspace per month, then all free web services suspend.

## Worked example

Workload: 4 vCPU / 8 GiB, 50 concurrent × 8 h/day × 22 days = **8,800 instance-hours**, 30% CPU, 50 GiB retained
state, 100 GiB egress. Modelled as 50 services on 4c-8g suspended between shifts (Render has no session API for
services; suspend/resume is your job).

Compute: 8,800 × $0.2397 = **$2,109.59** (CPU util irrelevant, allocated billing). State: 50 GiB of persistent disk ×
$0.25 = $12.50. Egress: plan-dependent.

| Regime | Feasible? | Monthly total |
|---|---|---|
| Hobby | No: 50 services > 25-service cap (also no autoscaling) | n/a |
| Pro, suspend between shifts | Yes | $25 + $2,109.59 + $12.50 + 75 GB × $0.15 ($11.25) = **$2,158.34** |
| Scale | Yes | $499 + $2,109.59 + $12.50 + $0 (1 TB incl.) = $2,621.09 |
| Scale + HIPAA (+20% on usage) | Yes | $499 + 1.2 × ($2,109.59 + $12.50) = $3,045.51 (bandwidth within 1 TB) |
| Pro, never suspended (always on) | Yes | $25 + 50 × $175 + $12.50 + $11.25 = **$8,798.75** |
| Cron 4c-8g (batch alternative) | Only for batch commands | $25 + 8,800 × $0.243 = $2,163.40 + extras |
| Workflows 4c-8g fixed | Batch tasks, ≤ 24 h runs | $25 + 8,800 × $1.00 = $8,825 + extras |
| Workflows flex | No: max 1 CPU / 4 GB | n/a |
| Sandboxes (EA) | No: 2 CPU / 4 GB limit, price unpublished | unknown |