Free
Token calculator

Claude Code vs Cursor: monthly cost for a solo developer

Claude Code starts at $20 a month on Claude Pro. Cursor plans run $8 to $200, but Claude models in Cursor bill at API list price on top (October 2026).

By · Updated

For a solo developer, Claude Code costs $20 a month on Claude Pro ($17 a month billed annually), or $100 or $200 a month on Max, with usage drawn from a plan allowance that Anthropic doesn’t publish as a number. Cursor’s individual plans cost $8 to $200 a month, but they include only Cursor’s own models (Grok and Composer). If you want Claude in Cursor, you pay the plan plus Claude tokens at API list price. All prices are as of October 2026.

How we built this: the plan prices, usage rules and model lists come from Anthropic’s and Cursor’s own pricing pages, docs and help center, read on October 25, 2026. Per-token model prices come from our AI model data, checked October 17, 2026, and they match what both companies list today. We haven’t used either tool for this comparison. The light, daily and heavy scenarios below are illustrative arithmetic on published prices, not measured usage and not real customers.

How much do Claude Code and Cursor cost per month?

Claude Code (on Claude plans)

Anthropic says “Claude Code is included in all paid plans.” The free Claude plan doesn’t include it.

Plan Price Usage, as Anthropic describes it
Free $0 No Claude Code
Pro $20/mo monthly, or $17/mo billed annually ($200 upfront) At least 5x Free per five-hour session
Max 5x $100/mo 5x Pro’s per-session usage
Max 20x $200/mo 20x Pro’s per-session usage
Team, Standard seat $25/seat/mo monthly, $20 annually More than Pro
Team, Premium seat $125/seat/mo monthly, $100 annually 5x a Standard seat

The Claude pricing page lists Max as “From $100” a month and says it’s billed monthly. Anthropic’s Max help article gives the two tiers as $100 and $200 a month.

You can also skip the subscription. Anthropic’s cost docs say plainly that “Claude Code charges by API token consumption” when you sign in with a Claude Console (API) account, so you pay per token with no plan fee.

Cursor

Plan Monthly Yearly What’s included
Free $0 – “Limited” agent and chat requests; no third-party models
Plus $8/mo $80/yr A weekly usage pool; Grok models only
Premium $30/mo $300/yr 4x Plus’s pool; third-party models at API list price
Super $100/mo $1,000/yr 15x Plus’s pool; priority model access at peak times
Ultra $200/mo $2,000/yr 50x Plus’s pool; earliest access to products and models
Teams $30, $100 or $200/seat/mo – Team management, SSO, usage analytics

Yearly prices come from Cursor’s pricing docs; the main pricing page shows monthly prices only. Cursor’s prices exclude tax. Cursor’s pricing page says subscriptions are sold through X Subscription Corp., and its docs point you to grok.com/dashboard/billing for usage credit settings.

What usage do you actually get?

Neither company publishes its included usage as a number of requests, tokens or dollars.

Claude: every plan has limits that reset on a rolling five-hour window, and paid plans add a weekly limit on top. Anthropic’s Pro and Max article gives no prompt count per five hours and no hours per week. The pricing page says there is no fixed message count. Claude chat and Claude Code share one pool: “all activity in both tools counts against the same limits.” Max plans also have a separate weekly limit for Anthropic’s top model, Fable. You can check what’s left with /status or /usage in Claude Code.

Cursor: paid plans include “a weekly pool of usage for Cursor models, shared across Grok and Cursor.” Cursor gives each plan’s pool as a multiple of Plus (4x, 15x, 50x) but never says how big Plus’s pool is. Cursor says the pool “refreshes weekly, so one heavy day doesn’t use up your month.” The Free plan is “Limited,” with no figure.

So for both tools, the only way to know if a plan covers your work is to watch the usage meter during your first week or two.

What happens when you run out?

This is where the two tools differ most.

Claude Code: when you hit a limit, you can wait for the reset, upgrade, or turn on usage credits. Anthropic’s help center says usage credits “are billed at standard API rates,” are prepaid (“Add funds”), can auto-reload, and can be capped with a monthly spend limit. They apply to Claude Code as well as chat. One catch: Max includes $100 or $200 a month of API credits for the Claude Platform, but Anthropic says those monthly API credits “can’t be used in Claude Code or the Claude apps.”

Cursor, own models: Grok and Composer usage beyond the weekly pool bills at Cursor’s per-token rates, paid from prepaid credit. You can top up by hand or turn on Auto Top-up, and set a “monthly additional usage limit.” Cursor says requests “are never downgraded in quality or speed” when the pool runs out.

Cursor, third-party models (Claude, GPT, Gemini and others): these are never in the pool. On Premium and above they bill “at API list price, paid from prepaid credit,” from the first request. Plus and Free can’t use them at all.

Which models can each one call?

Claude Code runs Anthropic’s Claude models. Its docs list the aliases sonnet, opus, haiku, fable and opusplan (Opus for planning, Sonnet for the edits). On Anthropic’s API these resolve to Sonnet 5.5, Opus 5.5 and Haiku 5.5. The docs say the default model on Pro, Max, Team, Enterprise and the API is Opus 5.5, and that no plan defaults to Fable. Depending on plan and seat, Fable usage can bill to usage credits instead of the plan allowance.

Cursor lists its own models (Grok 4.7, Grok 4.6, Grok 4.5 and Composer 2.5, which share the weekly pool) and a long list of third-party models from Anthropic, OpenAI, Google, Moonshot, Meta and Z.ai, among them Claude Sonnet 5.5, Claude Opus 5.5, Claude Haiku 5.5 and Claude Fable 5.1.

The per-token prices that matter for this comparison:

Model Input Cache read Output Where
Claude Sonnet 5.5 $2 $0.10 $10 Claude Code (API or usage credits); Cursor Premium+
Claude Opus 5.5 $4 $0.20 $20 Same
Claude Haiku 5.5 $0.10 $0.01 $0.50 Same (up to 100,000-token prompts)
Grok 4.7 $2 $0.50 $6 Cursor pool, then this rate
Composer 2.5 $0.50 $0.20 $2.50 Cursor pool, then this rate

Prices are USD per million tokens. The Claude and Grok figures are from our model data (checked October 17, 2026) and match the rates on Cursor’s models page; Composer 2.5’s is from Cursor’s models page. Writing to Anthropic’s prompt cache costs more than plain input: $2.50 per million for Sonnet 5.5 and $5 for Opus 5.5.

Cursor charges individual plans the same Claude rates Anthropic does. Teams and Enterprise pay an extra “Cursor Token Rate” of $0.25 per million tokens on eligible third-party models.

What does each cost at light, daily and heavy use?

These three scenarios are assumptions for illustration, not measurements. Coding agents resend the conversation with each step, so most input is cache reads; Anthropic’s cost docs show a sample session with 940,000 cache-read tokens against 1,200 fresh input tokens. Each scenario below is a month of tokens:

Scenario (illustrative) New input Cache reads Output
Light (a few sessions a week) 3M 30M 0.6M
Daily (most workdays) 15M 150M 3M
Heavy (long agent runs every day) 50M 500M 10M

We price new input at the base input rate and ignore the higher cache-write price, so real bills would run a little higher.

Worked example: the daily scenario on Claude Sonnet 5.5

  • New input: 15M × $2 = $30
  • Cache reads: 150M × $0.10 = $15
  • Output: 3M × $10 = $30
  • Total: $75 a month in tokens

The same tokens on Opus 5.5 cost twice as much: 15M × $4 + 150M × $0.20 + 3M × $20 = $60 + $30 + $60 = $150.

What that $75 of Sonnet tokens costs you through each route:

  • Claude Code on the API: $75, no plan fee.
  • Claude Code on Pro: $20, if the month fits inside Pro’s five-hour and weekly limits. Anything over can go to usage credits at the same API rates. Anthropic doesn’t publish the limits, so we can’t say whether it fits.
  • Cursor Premium with Claude Sonnet 5.5: $30 + $75 = $105, because Claude tokens aren’t in Cursor’s pool.
  • Cursor Premium with Composer 2.5 instead: $30, if the pool covers it. Past the pool, the same tokens at Composer’s rates are 15M × $0.50 + 150M × $0.20 + 3M × $2.50 = $7.50 + $30 + $7.50 = $45.

All three scenarios

Route Light Daily Heavy
Sonnet 5.5 tokens only (Claude Code on the API) $15 $75 $250
Opus 5.5 tokens only (Claude Code on the API) $30 $150 $500
Cursor Premium + Sonnet 5.5 tokens $45 $105 $280
Cursor Premium + Opus 5.5 tokens $60 $180 $530
Composer 2.5 tokens at overage rates (before any pool) $9 $45 $150

The arithmetic for the other cells: light Sonnet is 3M × $2 + 30M × $0.10 + 0.6M × $10 = $6 + $3 + $6 = $15; heavy Sonnet is 50M × $2 + 500M × $0.10 + 10M × $10 = $100 + $50 + $100 = $250. Opus doubles each. Light Composer is 3M × $0.50 + 30M × $0.20 + 0.6M × $2.50 = $1.50 + $6 + $1.50 = $9.

For a reference point, Anthropic’s own cost docs say that “across enterprise deployments, the average cost is around $13 per developer per active day and $150-250 per developer per month,” with 90% of users under $30 per active day. That’s Anthropic’s figure for API-billed teams, not a solo-developer number, but it sits inside the daily-to-heavy range above.

Try your own token counts in our AI token cost calculator.

What to watch for

  • Unpublished limits. Neither Claude’s five-hour and weekly limits nor Cursor’s weekly pool has a published size. A plan that covers one month may not cover a busier one.
  • Opus is Claude Code’s default. On an API key, Opus 5.5 costs twice Sonnet 5.5 per token. Anthropic’s docs suggest Sonnet for most coding and /model to switch.
  • A stray API key. Anthropic says that if ANTHROPIC_API_KEY is set, “Claude Code will use this API key for authentication instead of your Claude subscription,” so you’re billed per token even with Pro or Max.
  • Max’s API credits don’t cover Claude Code. The $100 or $200 a month of Claude Platform credits can’t be spent in Claude Code.
  • Claude in Cursor is never in the plan. Premium’s $30 doesn’t buy any Claude usage. Set a monthly additional usage limit before you pick a Claude model.
  • Thinking tokens. Anthropic’s docs say thinking tokens bill as output, and that Opus 5.5, Sonnet 5.5, Haiku 5.5 and the Fable models always use extended thinking. Lower the effort level for simple tasks.
  • Background use. Anthropic says Claude Code’s background jobs typically use under $0.04 per session, and agent teams use about 7x more tokens than a standard session when teammates run in plan mode.
  • Cancellation and renewals. Cursor says to cancel at least 24 hours before renewal. Claude’s Pro annual price is paid upfront ($200). Both companies say prices and plans can change.

Which should a solo developer choose?

  • You want Claude models at the lowest fixed price: Claude Pro at $20 a month ($17 billed annually) is the cheapest way to get Claude in a coding agent, if your usage fits its limits. Move to Max 5x ($100) or Max 20x ($200) if you hit limits often.
  • You code in bursts and want to pay only for what you use: Claude Code on an API key, with Sonnet 5.5 as the default model. In the light scenario that’s about $15 a month in tokens.
  • You want an AI editor and are happy with Grok or Composer: Cursor Plus ($8) or Premium ($30). Its own models come out of the weekly pool, and Composer 2.5’s overage rates are the lowest in the table above.
  • You want Cursor’s editor and Claude: budget the Cursor plan plus Claude tokens at API rates, or run Claude Code inside Cursor. Anthropic’s help center says Claude Code works in VS Code, Cursor and JetBrains IDEs on the same Pro or Max limits, which may cost less than paying for Claude tokens through Cursor.
  • You want models from several labs in one place: Cursor Premium and up can call Anthropic, OpenAI, Google and other models, each at its API list price. Claude Code runs Claude models.

For no-code and low-code builders like Lovable, Bolt and Replit, see our AI app builders hub and Lovable alternatives.

Sources