We cry about AI tools so you don't have to.

Pricing Watch

Figma Make Credit Pricing: What You Actually Pay Per Prompt

A Figma user spent $10.23 on one table edit — 341 credits at standard rates. Here's the actual cost breakdown.

figmafigma-makepricingai-tools
Disclosure: Tool Crier may earn a commission if you buy through links in this article. We paid for these tools ourselves or tested via free trials. Commissions don't influence what we write.

A Figma forum user reported spending $10.23 on a single table modification—341 credits at pay-as-you-go rates. That’s a design operation a human could do by hand in 45 seconds. We pulled the full pricing math so you can decide if Figma Make is a power tool or a money pit.

What a Credit Is, and When Figma Started Charging

According to Figma’s official update, Figma Make entered a new pricing era on March 18, 2026, when the company enforced seat-level credit limits across all plans. Before that date, credits were effectively unlimited for paid users. Now they’re metered and they run out fast.

A credit is a unit of consumption tied to any AI action in Figma: generating templates, removing backgrounds, vectorizing images, and running Figma Make agents. Every prompt burns a variable number depending on task complexity and model selection.

The Real Cost Per Action

Figma’s help documentation breaks down credit costs for individual features:

  • Removing a background: 1–5 credits
  • Adding interactions: 20 credits
  • Boosting image resolution: 5–10 credits
  • Font changes via Make: ~30+ credits
  • Adding interactivity via Make: ~75+ credits
  • Generating an app from scratch: ~100+ credits

That last one is the kicker. A full app prototype eats the same credits as roughly 3–4 font tweaks, yet both are billed as single-prompt operations. The system can’t predict consumption in advance. Figma’s documentation explicitly states that credit usage “varies based on task complexity, model selection, and context materials.” Translation: you won’t know the cost until after the agent runs and the credits are gone.

Forum users have reported numbers that reinforce this unpredictability. One designer posted that moving a toast message 50 pixels cost 64 credits. Another reported that a simple table modification consumed 341 credits. At $0.03 per credit on pay-as-you-go, that’s $10.23 per operation.

The Six Pricing Tiers (and What They Mean for Your Budget)

Figma offers credits through six distinct pricing paths:

PlanMonthly CreditsCost
Starter (free tier)500Included
Professional (full seat)3,000$16/month
Organization (full seat)3,500$55/month
Enterprise (full seat)4,250$90/month
Subscription add-on5,000$120/month
Pay-as-you-goUnlimited$0.03/credit

Here’s the practical translation: if you’re on a Professional plan at $16/month, your 3,000 monthly credits work out to approximately 50–70 Make prompts, depending on task weight. If you’re on Starter (500 credits), that’s closer to 8–10 prompts.

A Professional-plan user burning 341 credits per table edit would exhaust their monthly budget in fewer than 10 edits.

The 5.6x Seat Math (and Why Finance Teams Are Furious)

The Figma forum lit up when users discovered the pricing arbitrage:

  • Adding a full Professional seat: $16/month = approximately $0.005 per credit (dividing seat cost by included 3,000 credits)
  • Buying additional credits at pay-as-you-go: $0.03 per credit

That’s a 5.6x premium for standalone credits compared to seat-bundled credits. In other words, if your team exhausts Make’s monthly allowance, the rational move isn’t to buy more credits—it’s to add another user seat, even if the extra person doesn’t exist. You’ll pay $16 for 3,000 credits instead of $150 for 5,000 credits at the subscription tier.

Figma’s math hasn’t gone unnoticed. Users are openly questioning whether this pricing reflects the actual cost of API inference or whether it’s deliberate friction to push teams toward seat-based licensing.

Model Choice Costs More Than You’d Expect

Figma Make lets you pick your underlying model: Gemini 3 Flash, Gemini 3.1 Pro, Claude Sonnet 4.6, or Claude Opus 4.7. Here’s the catch: Claude Opus 4.7 consumes substantially more credits per task than other available options.

Figma hasn’t published exact cost differentials between models (they say “check your credit history to see the difference”), but the pattern is clear: stronger models burn more credits. If you’re running Make in a budget-constrained context, selecting Gemini 3 Flash or Claude Sonnet 4.6 will stretch your allowance further than defaulting to Opus.

This is the one dial you can actually turn. Everything else—task complexity, context size, prompt phrasing—is outside your control once you hit send.

So: What Should You Actually Do?

If you’re on a Professional plan and Make is core to your workflow: add a second seat (or a Dev seat at $55/month for Organization). The math favors it. You’ll get 3,000–3,500 more credits for either $16 or $55 less than buying credits outright.

Starter and free-plan users are effectively locked out of Make for now. At 500 credits per month and 30+ credits per action, you’ll hit the wall in a handful of edits.

Experimental or occasional users: pay-as-you-go is fine, but watch the balance. One table edit doesn’t sting. Ten of them do.

If you’re evaluating whether Make is worth it versus other vibe-coding tools: note that Figma’s credit model is opaque by design, unlike Replit’s transparent effort-based billing or v0’s upfront token pricing. You won’t know your true cost until after you’ve shipped. Compare this friction to tools like Lovable or Bolt, where credit consumption is at least predictable.

Figma Make is powerful. It’s also expensive in ways that don’t scale linearly with your team size or usage. Until Figma decouples credits from seats or publishes per-action cost guarantees, the safest bet is to treat Make as a premium feature, not a daily workhorse. If you’re a heavy user, the seat arbitrage is there—use it.

The pricing is baroque enough that it deserves a translator. That’s the cry.

Sources:

← More Pricing Watchs

What we don't know is documented at the end of this article. We update when we learn more.