Analysis · April 2026

Claude Design pricing: what does it really cost?

A clear cost analysis of Anthropic's new AI design tool, compared with Figma Make, Lovable, v0, Bolt.new and Cursor. Written for Dutch organisations that want to make a sensible choice without the hype.

In shortWhat to expect from this document

Claude Design launched on 17 April 2026 as an Anthropic Labs research preview and runs on Claude Opus 4.7. Access starts with the Pro subscription at roughly €18.50 per month excluding VAT, but the real cost is often higher than the website price suggests.

Claude Design is metered separately from your regular Claude conversations and Claude Code, with its own weekly usage allowance that is still in the trial period and not yet published as fixed numbers. Anyone who wants to iterate regularly will quickly move towards Max 5x, around €92 per month, or Max 20x, around €185 per month, or pay for extra usage at the standard Opus 4.7 API rate. This document lays out the facts, compares the familiar alternatives and gives a sober conclusion on when Claude Design is worth it.

About the prices in this document

Anthropic invoices in US dollars and does not publish a separate EUR price list. The euro amounts in this document are conversions based on an exchange rate of approximately €1 = $1.08, reference date 27 April 2026, excluding VAT. For Dutch private customers, 21% VAT is added to the USD price. For business customers with a valid VAT number, the reverse charge mechanism usually applies through Anthropic Ireland.

01How Claude Design's cost model works

Claude Design sits inside your existing Claude subscription, but Anthropic deliberately separates it into its own meter. In plain English, the official documentation comes down to this:

Claude Design is priced and metered separately from the rest of Claude. It has its own usage tracking, its own usage allowance and, for subscriptions, its own weekly limits next to your existing limits for conversations or Claude Code.

Source: support.claude.com, Claude Design subscription, usage and pricing

Three consequences are worth noting.

First, your Pro, Max or Team limit for regular conversations does not determine how much Claude Design you get. It is a separate allowance, per user, that resets every 7 days.

Second, this allowance is per person and is not shared across an organisation. You cannot create a shared team pool where one designer uses a lot and everyone else uses little.

Third, Anthropic still describes these limits as trial-period limits that may change. What fits within your subscription today may be different in two weeks. That is not a detail. Anyone building production work on this takes on risk.

The calculation rules in practice

What is firm and publicly confirmed:

  • Claude Design runs on Claude Opus 4.7, Anthropic's most expensive vision model.
  • If you exceed your weekly usage allowance, you can enable extra usage. That is charged at standard API rates: $5 per million input tokens and $25 per million output tokens.
  • Larger images consume tokens faster. Opus 4.7 supports images up to 2576 pixels, and the larger your input, the larger the token bill.
  • For consumption-based Enterprise, Anthropic says Claude Design usage is charged at the standard API rates in the existing agreement. A one-off Enterprise credit of around 20 typical tasks per user applies as an introductory allowance until 17 July 2026.
What Anthropic does not publish

A fixed number of credits, tasks or designs per Pro, Max or Team subscription. The phrase "typical tasks" only appears in the Enterprise context. For consumer plans, the wording remains: weekly usage allowance, trial period, subject to change.

02Which subscriptions provide access?

Source: support.claude.com, Anthropic Labs admin guide, April 2026.

Subscription Claude Design access Intended for
FreeNoNot available
ProYesExploration and occasional use
Max 5xYesRegular use, product managers and developers with recurring sample designs
Max 20xYesIntensive users, designers and creatives
Team StandardYesTeam-level exploration
Team PremiumYesIntensive users, including Claude Code
Enterprise, old modelYesStandard and Premium licences by role
Enterprise, consumption-basedYesUsage at the API rate in the existing agreement

For Enterprise, Claude Design is off by default. An administrator must enable it in the organisation settings. In addition, audit logs and usage tracking for Claude Design are currently not available, and data residency in a specific region is not supported. For highly regulated sectors, that is a serious concern.

03What does each subscription cost?

All prices were taken from claude.com/pricing on 27 April 2026. Anthropic invoices in USD. The EUR amounts below are indicative conversions excluding VAT.

Individual plans

Plan EUR / month
indicative, excl. VAT
EUR / year
indicative, excl. VAT
USD / month USD / year
Free€0€0$0$0
Pro, monthly~€18.50~€222$20$240
Pro, annual upfront~€15.75~€189$17 effective$200
Max 5x~€92.60~€1,111$100$1,200
Max 20x~€185.20~€2,222$200$2,400

Team and Enterprise

Plan EUR / seat / month
indicative, excl. VAT
USD / seat / month Conditions
Team Standard, annual~€18.50$20Minimum 5 seats
Team Standard, monthly~€23.15$25Minimum 5 seats
Team Premium, annual~€92.60$1005x more usage than Standard
Team Premium, monthly~€115.75$125Includes Claude Code
EnterpriseOn request$20 per seat + token usageUsually minimum 50 seats, annual contract

Important to know about Team: Claude Code is only included in Team Premium, not in Team Standard. Anyone who wants to make real use of the hand-off to Claude Code will quickly end up on the higher tier.

04Practical cost examples

Because Anthropic does not publish a fixed number of designs per subscription, we use the only firm public reference point available: one Enterprise credit covers around 20 typical tasks. That is the unit Anthropic itself uses. We calculate from there, while recognising that a finished design often requires more than one task.

Example 1: creating 10 worked-out designs

In Anthropic's own example, one design takes one to a few tasks plus several adjustment rounds and an export file. Realistically, based on early user experiences such as the Brilliant case, where the most complex pages took 2 tasks, you should assume 3 to 8 tasks per finished design. For 10 designs, that means 30 to 80 tasks.

Anthropic does not share exact token counts per task, but as a guideline for a task with image input and design output, you can estimate roughly 50,000 to 200,000 total tokens per task. At Opus 4.7 rates, $5 input and $25 output per million, that puts you between $0.40 and $2.50 per task.

Scenario Number of tasks Estimated extra API cost Indicative in EUR
10 simple designs, little iteration30$12 to $75~€11 to ~€69
10 moderately complex designs60$24 to $150~€22 to ~€139
10 intensive designs with many images80$32 to $200~€30 to ~€185

If these tasks fit within your weekly usage allowance on Pro or Max, they do not cost anything extra. But as Anthropic itself says, the allowance is variable and can change.

Example 2: how many designs per month on Pro and Max?

The honest answer based on publicly available information: Anthropic does not publish this as fixed numbers. The official guidance is qualitative:

  • Pro: quick explorations and one-off use.
  • Max 5x: regular use by product managers and developers who frequently create sample designs.
  • Max 20x: intensive use by designers and creatives.

Translated into a practical expectation:

Plan What you can realistically do per week
our estimate based on Anthropic's wording
ProA few exploratory sessions, 1 to 3 finished designs per week before you run into limits
Max 5xRegular output, 5 to 15 designs per week, with room to iterate
Max 20xDaily use at design-team level, potentially 20 or more designs per week

These numbers are not confirmed by Anthropic. They are derived from the "1x, 5x, 20x" naming and the product positioning. Anyone relying on this for budgeting should calibrate usage in the first week and adjust the subscription accordingly.

Example 3: cost over a full year

Package EUR / year
indicative, excl. VAT
USD / year What you get
Pro annual~€189$200Access for one user, light projects
Max 5x~€1,111$1,200One intensive user, ample room for sample designs
Max 20x~€2,222$2,400One full-time designer
Team Standard, 5 seats annual~€1,111$1,200Small team, light exploration
Team Premium, 5 seats annual~€5,556$6,000Small design team with SSO and admin capabilities

A Dutch design team of five people using Claude Design seriously will realistically land between €5,500 and €12,000 per year, excluding any extra API costs. If you are only experimenting, €189 for annual Pro is enough.

05Comparison with other AI design tools

All prices were verified on 27 April 2026 on the suppliers' official pricing pages. Units differ by platform, so first the table, then the interpretation. All EUR amounts are indicative conversions excluding VAT.

Pricing table per tool, entry-level paid version

Tool Free version First paid version Unit
Claude Design (Pro) Not available ~€18.50/month ($20)
~€15.75/month annually ($17)
Weekly usage allowance, trial period, not publicly specified
Figma Make (Professional) 500 AI credits/month on free Starter ~€18.50/month ($20) per full licence 3,000 AI credits/month, ~30-50 Make generations
Lovable (Pro) 5 credits/day, max 30/month ~€23.15/month ($25) 100 credits + 5/day extra, ~80-150 interactions
v0 by Vercel (Premium) $5 in credit ~€18.50/month ($20) $20 in token credit, variable
Bolt.new (Pro) 1M tokens/month, cap 300K/day ~€23.15/month ($25) 10M tokens, ~100-200 medium-sized tasks
Cursor (Pro) Limited Agent requests ~€18.50/month ($20) $20 model credit + unlimited automatic mode

What this really means

Claude Design is the least transparent in its price per unit. Lovable, Cursor and Bolt at least give you a numerical lower bound, such as credits, requests or tokens. For Pro and Max, Anthropic only says: weekly usage allowance, variable, trial period. For Enterprise, Anthropic names one reference point: 20 typical tasks per credit. That is unusually little certainty for a design tool aimed at production output.

Figma Make sits in a different category. It is integrated into a design platform that is already widely used. For 30 to 50 generations per month, you pay $20 per full licence. Once your credit is used up, it simply reaches zero, after which you can buy more at $0.03 per credit. Pay-as-you-go is active on Org and Enterprise, and on Professional from May 2026. Clear and understandable.

Lovable and Bolt focus on complete applications rather than designs alone. If you only want a prototype without server logic, you are paying there for functionality that Claude Design does not even provide: Claude Design does not generate the server side, only the visual hand-off to Claude Code.

v0 and Cursor are primarily tools for developers. For a Dutch marketer or designer who "just wants to see something quickly", they are less suitable than Claude Design or Figma Make.

Price per result, rough ranking

Based on publicly available figures, using Pro as the reference for the entry-level paid version:

Rank Tool Indicative cost per result
1Cursor Pro~$0.09 per Claude Sonnet request, code only
2Bolt Pro$0.13 to $0.17 per medium-sized task
3Lovable Pro~$0.25 per credit
4Figma Make Professional~$0.50 per Make generation
5Claude Design ProCannot be determined from public information

That final point is the problem. Anthropic explains this by saying it is a research preview and that the usage allowance is deliberately qualitative. For early users, that may be acceptable. For anyone who has to plan a budget, it is a familiar Anthropic pattern: you pay a high price for uncertainty, and you have to calibrate usage yourself.

06Hidden costs and points to watch

Six points that are not on the sticker price but can still show up on your bill.

Extra usage may be enabled without you noticing

Anthropic offers extra usage above your weekly allowance, charged at API rates. On Team and Enterprise, administrators can manage this through spending limits. On Pro and Max, you are responsible yourself. A review on The New Stack reported confusion: "I was out of tokens and started consuming extra purchased tokens, even though I had not used my full session allowance". The signal is clear: monitor your usage.

The hand-off to Claude Code uses your Claude Code allowance, not your Claude Design allowance

You prepare the hand-off bundle in Claude Design, but the actual implementation runs in Claude Code and consumes tokens there. For one finished product, you are paying through two separate meters.

Pro gets Claude Code, Team Standard does not

A team that wants Claude Design at Pro level for designers and Claude Code for developers must choose Premium within Team, around €92 per seat per month when billed annually. Otherwise, Standard lacks Claude Code access.

No audit logs or usage tracking for Claude Design

Enterprise has an enablement switch, but no detailed reporting per user. Anyone who needs to allocate costs per department lacks information here.

No fixed data residency by region

For Dutch organisations with strict GDPR or sector requirements, such as healthcare, government or finance, Claude Design is currently not an option for production work with confidential data. Anthropic explicitly confirms that data residency in a specific region is not supported.

Beta status means prices and limits may change

Read Anthropic's caveat literally: subject to change. That is realistic for a product that is 10 days old, but it belongs in your risk assessment.

07When is Claude Design worth the price?

Three situations where Claude Design may be a sensible choice.

You already use Claude Pro or Max every day

Then Claude Design is effectively included within your weekly allowance. For light exploration, it costs you nothing extra. Start there.

You have an existing codebase and want to derive the design style from it

Claude Design reads your React components or existing design files and extracts colours, typography and components from them. That puts it ahead of v0 and Lovable in this specific workflow. If you can use that advantage, you gain speed.

You want to move smoothly from prototype to production through Claude Code

The hand-off bundle is the clearest differentiator. An early customer, Brilliant, reported 2 tasks instead of more than 20 in other tools. That is a specific claim, but there is a confirmed difference in workflow.

Three situations where Claude Design is not the right choice.

You need predictable costs

The opaque weekly allowance makes Claude Design hard to budget. Lovable, Bolt and Figma Make are more transparent.

You work with confidential or regulated data

No audit logs, no fixed data residency by region and no usage tracking. Wait until this is mature enough for business use.

You already have a productive Figma team

Figma Make fits directly into your existing workflow and, at the time of writing, is included in every paid Figma licence. For incremental improvement, you do not need to buy a second platform.

08Conclusion: the sober cost reality

Claude Design positions itself in the higher segment and behaves accordingly. The entry price of around €18.50 per month is misleading, because the real usage limits and extra API-rate usage can make your bill higher than a Bolt or Lovable subscription at comparable intensity. At the same time, Claude Design is the only tool that offers a closed chain from extracting design style from code to production implementation with Claude Code. For teams that need that chain, this is a real advantage.

Three recommendations

1. Start with your existing Pro or Max subscription. Use the research preview for what it is: an exploration. Track your usage during the first two weeks. Adjust your subscription based on what you actually measure.

2. For a productive solo designer, expect a move to Max 5x, around €92 per month, or Max 20x, around €185 per month. For a team of five, Team Premium at around €5,556 per year is realistic.

3. Wait before using it for confidential projects until Anthropic provides fixed data residency by region, audit logs and usage tracking for Claude Design. Until then, it is an excellent exploration tool, but not a mature business tool.

SourcesVerified on 27 April 2026

More control over AI tools?

Train your team to work responsibly with Claude and other AI tools

Mellaart Trainingen helps Dutch organisations use AI practically, safely and purposefully. Online or on location, always tailored to your way of working.