Sanity vs Payload CMS vs Hygraph: Pricing Comparison 2026

Sanity, Payload CMS, and Hygraph each target different budgets and team sizes — and the pricing structures are different enough that a direct comparison saves real money at decision time. This post puts all three side by side with concrete numbers, not marketing copy.

Why this three-way matters

Most pricing comparisons pit Sanity against Contentful or Storyblok. Those matchups get covered. What gets missed is the triangle formed by Sanity (API-first, usage-based), Payload (self-hosted open-source with optional cloud), and Hygraph (GraphQL-native, seat-gated). They compete for the same Next.js project budget at the $0–$1,000/month band, but the upgrade triggers are completely different — and picking the wrong one means an unexpected bill at the worst possible moment.

Free tier limits at a glance

Payload's "free" tier is really just the open-source licence — you still pay for compute, Postgres, and storage wherever you deploy it.

Sanity pricing in 2026

Sanity runs on usage, not seats. The Free plan gives you 500K CDN requests and 10 GB of bandwidth per month. The Growth plan starts at $15/project/month (billed annually) and raises limits to 1M CDN requests and 100 GB bandwidth. Above that, you pay overage:

• CDN requests: ~$0.008 per 1K over limit
• Bandwidth: ~$0.10/GB over limit
• API mutations (writes): counted against a separate quota

The real upgrade trigger on Sanity is almost never seats — it's bandwidth on image-heavy sites. A marketing site pulling 50 Sanity-hosted images per page view at 200–400 KB each hits 10 GB faster than you'd expect. Run the numbers before launch.

Enterprise pricing is custom and gated behind a call. Public signals suggest $1,000+/month for high-volume API tiers with SLAs.

Payload CMS pricing in 2026

Payload's open-source version is MIT-licenced and genuinely free to self-host. The real cost is infrastructure:

• Railway (hobby): ~$5–20/month for a small Node + Postgres instance
• Railway (production): $50–150/month depending on memory and storage
• Vercel + Neon: $0–25/month at small scale, but Payload's local API won't work in an edge runtime — you need a dedicated Node process
• Payload Cloud (beta → GA in 2026): Starter at $40/month, Growth at $100/month, with dedicated environments and background jobs included

Payload Cloud is the path-of-least-resistance managed option. The Growth tier at $100/month includes 3 environments (local, staging, prod), custom domains, and email via Resend. There's no per-seat cost — you control access in code via access control functions.

The upgrade trigger on Payload is almost always compute: scheduled jobs, background processing for uploads, or a media library that's grown beyond what a small Railway instance can handle without spiking memory.

Hygraph pricing in 2026

Hygraph's free tier is more restrictive than it looks. 3 seats and 1 project sounds fine until you add a staging environment — that's environment #2, which is included on Free but with reduced rate limits. The paid plans:

• Professional: $399/month (billed annually — yes, that's not a typo)
• Scale: $799/month
• Enterprise: custom

Hygraph also charges per seat above the included count. Professional includes 10 seats; each additional seat is ~$25/month.

The Professional plan is a significant cliff. Most teams hit it when they need:

• More than 3 editors (Free limit)
• Webhooks at production rate limits
• Per-environment content permissions
• Content federation (querying external APIs from within Hygraph)

Hygraph's GraphQL-native approach and content federation feature are genuinely differentiated, but $399/month is a hard number to justify for a small team when Sanity Growth at $15 covers most of the same content modelling surface.

Side-by-side cost at three team sizes

What actually triggers an upgrade on each platform

Sanity: bandwidth overages from media-heavy pages. A Sanity image URL without ?w=800&auto=format serves the original. One unoptimised image fetch per visitor at scale blows through 10 GB quickly. Fix: always use the @sanity/image-url builder with explicit width and format parameters, and lean on next/image to gate what gets requested.

Payload: compute. The Node process doing image processing, running hooks, and serving the Admin UI can't live in a serverless function. You need persistent compute, and that costs money whether you use Payload Cloud or your own VPS. Budget $50–100/month minimum for a production-grade Postgres + Node setup.

Hygraph: seats and the $399 cliff. If a client has 4+ editors who need write access, you're paying $399/month. There's no middle tier between Free and Professional. This is Hygraph's most common gotcha — plan for it in the proposal stage.

Which one fits which project

Sanity Free-to-Growth works best for marketing sites and blogs where content editors are few (under 10) but API traffic is high. The seat model never bites you; the bandwidth model might.

Payload self-hosted is the right call when you want zero platform lock-in and your team can manage infrastructure. It's also the only option here with a fully custom access control system in code — no permissions UI to fight with.

Hygraph makes sense when the content federation feature is load-bearing — pulling from multiple APIs into a single GraphQL layer that editors see as one content surface. At $399/month you need that feature to justify the cost over Sanity.