Render's pricing page separates the workspace from the work. Hobby is $0 per month plus compute, Pro is $25 per month plus compute, Scale is $499 per month plus compute and Enterprise is custom. Compute is then priced per instance type, and the page publishes the whole ladder: 512 MB RAM with less than one CPU for $7 per month, 2 GB with one CPU at $25, 4 GB with two CPUs at $85, 8 GB with two CPUs at $135, 16 GB with two CPUs at $200, 8 GB with four CPUs at $175, 16 GB with four CPUs at $225 and 32 GB with four CPUs at $350. A free instance of 512 MB is listed with limitations applying, and persistent SSD storage is $0.25 per GB per month.
TL;DR
Both platforms have a free tier and a $20-something paid tier, but they charge for completely different things. Render bills a workspace fee, Hobby $0, Pro $25 per month and Scale $499 per month, and then bills compute per instance, from $7 per month for 512 MB of RAM up to $350 per month for 32 GB, with block storage at $0.25 per GB per month. Vercel bills $0 for Hobby and $20 per month for Pro, which includes $20 of usage credit, and then meters almost everything else: bandwidth, firewall requests at $0.50 per million, ISR reads at $0.40 per million, Blob storage at $0.023 per GB and image transformations at $0.05 per thousand. Render suits long-running services and databases that need predictable instance pricing; Vercel suits front-end and function workloads where traffic, not uptime, drives the bill.
Vercel's page works the other way around. Hobby is $0 per month and Pro is $20 per month, and Pro includes $20 of usage credit, a flat-rate CDN with traffic spike protection and advanced spend management. Everything above that is metered in units: firewall rate-limit requests at $0.50 per million after the first million, BotID deep analysis at $1 per 1,000 checks, incremental static regeneration reads at $0.40 per million after the first million and writes at $4 per million after the first 200,000, Blob storage at $0.023 per GB with simple operations at $0.40 per million and advanced operations at $5 per million, image optimisation transformations at $0.05 per thousand after 5,000 free, and bulk redirects at $50 per month per 25,000.
The two models reward different workloads. A Render service runs continuously and is billed whether or not anyone visits, which is what you want for an API, a database, a worker or a cron job, and it means the monthly figure is knowable in advance. A Vercel deployment is billed by what it serves, which is cheaper when traffic is bursty and expensive when it is heavy, and its 2 included microfrontend projects and $250 per additional project price shows how far the usage-model reaches. Render also advertises 7 million plus builders, up to $10K in migration credits and up to $100K in credits through its startup programme, while Vercel's page carries regional pricing for compute.
| Feature | Render | Vercel |
|---|---|---|
| Free tier | Hobby workspace at $0/month plus compute; a $0/month 512 MB instance is listed with limitations applying | Hobby at $0/month with automatic CI/CD, Web Application Firewall, global CDN, DDoS mitigation and traffic insights |
| Entry paid tier | Pro workspace $25/month plus compute | Pro $20/month including $20 of usage credit, flat-rate CDN and spend management |
| Top published tier | Scale $499/month plus compute; Enterprise at custom pricing with dedicated support and uptime SLAs | Enterprise at custom pricing with custom OWASP rulesets, custom bot management and advanced security |
| What the plan price buys | Workspace features; capacity is bought separately as instances | A platform fee plus $20 of usage allowance; most additional consumption is metered |
| Cheapest compute option | $7/month for 512 MB RAM with less than one CPU, and a $0/month 512 MB instance with limitations | Not sold per instance; functions are billed through the usage meters |
| Largest published compute option | $350/month for 32 GB RAM and four CPUs, with a $300/month 16 GB and eight-CPU option also listed | Compute is regional and usage-priced rather than sized per instance |
| Storage | Persistent SSDs at $0.25 per GB per month | Blob storage at $0.023 per GB; simple operations $0.40 per million after 10K, advanced operations $5 per million after 2K |
| Traffic-related charges | Included in the instance price rather than metered per request | Firewall rate-limit requests $0.50 per million after 1M; ISR reads $0.40 per million after 1M; data transfer 10 GB/month included, then CDN-based |
| Image and caching | Not metered separately on the compute line | Image transformations $0.05 per thousand after 5K; cache reads $0.40 per million after 300K; cache writes $4 per million after 100K |
| Enterprise-only capabilities | Dedicated support and uptime SLAs; migration credits up to $10K and up to $100K in startup credits | Custom OWASP Core Ruleset, custom bot management, additional microfrontend projects at $250 each beyond 2 included |
| Best fit | Always-on services, APIs, workers and cron jobs where capacity is known | Front-end and function workloads where traffic shape varies and edge delivery matters |
Pick Render if
Render's structure suits anything that runs continuously. The workspace fee is small, $0 or $25 per month, and the cost of the actual service is a published instance price you can pick from a table, from $7 per month for 512 MB to $350 per month for 32 GB with four CPUs. That makes a monthly bill predictable in a way meters never are: an API, a queue worker, a database and a cron job each get a line, and none of them changes because traffic doubled for a day. Persistent storage at $0.25 per GB per month is also priced openly. The cost of that predictability is paying for idle capacity and choosing from a long instance menu, plus Scale at $499 per month before you reach the managed support commitments.
Pick Vercel if
Vercel makes sense when your workload is front-end delivery and functions that scale with visitors rather than clock time. Pro is $20 per month and hands back $20 of usage credit, with a flat-rate CDN that absorbs traffic spikes and spend management to stop surprises, and the metered rates are published to four decimal places: $0.023 per GB of Blob storage, $0.40 per million ISR reads, $0.05 per thousand image transformations. Hobby at $0 is a genuine start for a personal project. Budget attention goes to the small thresholds and the steps: bulk redirects after 1K per project at $50 per month per 25,000, microfrontends at $250 per additional project beyond 2, and enterprise-grade security features that never appear on a published price.
Neither Render nor Vercel publishes third-party review data we are willing to restate as fact, and we did not run either platform ourselves. Every figure here is taken from the two vendors' own pricing pages as published when this comparison was written. Usage meters change quietly, so if a rate has moved since, tell us and we will re-check it.
It depends on what you are running. Vercel's paid tier is $20 per month and Render's is $25, both plus actual usage. Render then charges per instance, from $7 per month for 512 MB, while Vercel charges per unit consumed. A single small always-on service is cheaper on Render; a static front end with steady traffic is cheaper on Vercel's Hobby plan at $0.
Yes, and both are permanent rather than trials. Render's Hobby workspace is $0 per month plus compute, with a $0/month 512 MB instance listed with limitations applying. Vercel's Hobby is $0 per month with automatic CI/CD, WAF, global CDN, DDoS mitigation and traffic insights.
Pro includes $20 of usage credit each month alongside the $20 plan price, which is consumed by the metered services such as firewall requests, ISR reads and writes, Blob storage and image transformations. Flat-rate CDN with traffic spike protection and advanced spend management are included rather than metered separately.
Render, structurally. Its page sells instances by RAM and CPU at published monthly prices, from $7 to $350 per month, plus persistent SSDs at $0.25 per GB per month, which is the shape a database or a worker needs. Vercel's pricing page meters storage and functions rather than renting persistent compute.
A 512 MB instance with less than one CPU is listed at $0 per month with limitations, and 512 MB with 0.5 CPU is $7 per month. At the top, 32 GB with four CPUs is $350 per month, with a 16 GB, eight-CPU option listed at $300.
Vercel's page is more specific: custom OWASP Core Rulesets, custom bot management, BotID deep analysis at $1 per 1,000 checks and tenant isolation are listed, with several held for Enterprise. Render holds dedicated support and uptime SLAs for Enterprise at custom pricing rather than publishing security capability levels.
Spot an outdated price or feature? Let us know and we'll verify it.