[{"data":1,"prerenderedAt":53},["ShallowReactive",2],{"blog-claude-code-pro-vs-max":3,"blog-related-claude-code-pro-vs-max":36},{"slug":4,"title":5,"description":6,"category":7,"categoryId":8,"keyword":9,"date":10,"minutes":11,"body":12,"html":13,"headings":14},"claude-code-pro-vs-max","Claude Pro or Max for Claude Code: which plan fits you?","Claude Pro or Max for Claude Code? How to decide from your own usage: how often you hit the limit, how many sessions you run, and what else you could do.","Usage & limits","usage-limits","claude pro vs max claude code","2026-11-17",5,"Pick Claude Pro if you use Claude Code in short, focused sessions and rarely see a limit message. Pick Max if you hit the 5-hour or weekly limit regularly, run several sessions in parallel, or use Claude Code as your main way of working. The deciding factor is not features but how often you run out, so measure that before paying more. Quotas and prices change, so this article gives none; check the plans page on claude.com.\n\n## What is actually different between Pro and Max for Claude Code?\n\nBoth plans give you Claude Code with the same tools, commands and workflows. What differs is the size of the allowance: Anthropic's help center describes a five-hour session limit and a weekly limit on both, with Max offering more room than Pro. Higher tiers also bring some model-specific allowances, and which models are included in the plan or billed to usage credits can differ by plan. The `\u002Fmodel` picker marks models that \"require usage credits\", and extended context for some models depends on plan too. Look at the picker on your own account rather than a blog post (including this one) for the current state.\n\nLimits are shared across Claude chat and Claude Code. If you also use claude.ai heavily, that is part of the same pool, which is an argument for Max even if your Claude Code use alone is modest.\n\n## How do I know whether I need Max?\n\nUse data, not feeling. For two weeks, note:\n\n1. **How often you hit a limit.** A session limit once a week that resets while you eat lunch is not a reason to upgrade. A limit every afternoon is.\n2. **What was running when you hit it.** `\u002Fusage` breaks recent usage down by skills, subagents, plugins and MCP servers, and recommends changes. If one habit is behind a big share (a loop, an always-on subagent, a giant context), fix that first.\n3. **Whether you hit the weekly limit.** The 5-hour window is an annoyance; the weekly one stops you for days. Hitting it more than once in a while is the clearest upgrade signal.\n4. **How many sessions you run at once.** Parallel sessions multiply usage. See [running several Claude Code sessions](\u002Fblog\u002Frun-multiple-claude-code-sessions). Two or three on a big codebase is a different load from one.\n5. **How much you wait.** Count the hours per week that Claude Code was stopped and you had work for it. Price that against the plan difference.\n\n## Can I cut usage before upgrading?\n\nOften, yes, and it is free. Most heavy usage comes from context size, not from the number of prompts: the whole conversation is re-sent with every request. Before you upgrade:\n\n- `\u002Fclear` between unrelated tasks and `\u002Fcompact \u003Cfocus>` in long ones ([the context window, explained](\u002Fblog\u002Fclaude-code-context-window)).\n- Lower the effort with `\u002Feffort` on simple tasks, and use a lighter model for routine edits and subagents.\n- Keep CLAUDE.md short and move rarely needed instructions to skills.\n- Use plan mode before large changes, so you do not pay for a wrong implementation.\n- Stop loops and background agents you forgot about.\n\nThe full list is in [Claude Code usage limits](\u002Fblog\u002Fclaude-code-usage-limits). If you do all of this and still hit the weekly limit, the upgrade is justified.\n\n## What are the options besides changing plan?\n\nYou are not choosing between only Pro and Max.\n\n- **Usage credits.** You can turn them on to keep working past a limit; they bill at API rates, with a monthly spend limit you set. `\u002Fusage-credits` opens the settings. Good for occasional overflow, expensive as a habit. If you overflow every week, a plan is likely cheaper. See [what your usage would cost at API prices](\u002Fblog\u002Fclaude-code-token-costs).\n- **Team or Enterprise seats.** If you work for a company, a seat has its own allowance, and admins can add usage credits and spend limits.\n- **A second account that is yours.** A personal plan plus a work seat, each with its own limits, is legitimate when they are for different purposes. See [using several Claude accounts with Claude Code](\u002Fblog\u002Fclaude-code-multiple-accounts). Do not use extra accounts as a trick to dodge limits; read Anthropic's terms first.\n- **API billing (Console).** Pay per token, no windows. Predictable only if you watch it. For heavy interactive use, compare it with a plan using your own token numbers (see the token costs article above); it is also a different login mode from a subscription.\n- **Waiting.** With automatic continue in recent Claude Code versions, a session stopped at a limit picks the task up after the reset if you keep it open.\n\n## Which plan for which kind of user?\n\n| If you... | Lean toward |\n| --- | --- |\n| Use Claude Code a few hours a week, one session at a time | Pro |\n| Mostly use chat and occasionally ask Claude Code for a fix | Pro |\n| Hit the 5-hour limit most workdays | Max |\n| Hit the weekly limit | Max, after trimming context habits |\n| Run parallel sessions, worktrees or scheduled agents | Max |\n| Work in a company with a Claude plan | A team seat, plus your own plan for side work |\n| Need a pay-per-token bill for accounting reasons | API (Console), knowingly |\n\nTreat the first month on a higher plan as a test. You can compare how often you were blocked, and downgrade if the answer is never.\n\n## A trap to avoid: the API key\n\nIf `ANTHROPIC_API_KEY` is set in your shell, Claude Code uses it (once you approve it) and bills the API, not your plan. People pay for Max and then burn API credits without knowing. Run `\u002Fstatus` to see which credential is active, and `unset ANTHROPIC_API_KEY` if you want the subscription.\n\n## In Relais\n\nRelais drives the Claude Code login already on your machine, with Pro, Max or a team seat; it has no plan of its own for Claude usage and needs no API key. Everything it runs counts toward your plan's limits, and your 5-hour usage shows under your account in the sidebar. Running several sessions and agents at once is exactly what makes a plan feel small, so Relais lets agents stop starting new work past a share of your plan, and carries a session over to your next account when one runs out. See [pricing](\u002Fpricing) for how Relais and your Claude plan are billed separately.","\u003Cp>Pick Claude Pro if you use Claude Code in short, focused sessions and rarely see a limit message. Pick Max if you hit the 5-hour or weekly limit regularly, run several sessions in parallel, or use Claude Code as your main way of working. The deciding factor is not features but how often you run out, so measure that before paying more. Quotas and prices change, so this article gives none; check the plans page on claude.com.\u003C\u002Fp>\n\u003Ch2 id=\"what-is-actually-different-between-pro-and-max-for-claude-code\">\u003Ca class=\"anchor\" href=\"#what-is-actually-different-between-pro-and-max-for-claude-code\" aria-label=\"Link to this section\">#\u003C\u002Fa>What is actually different between Pro and Max for Claude Code?\u003C\u002Fh2>\n\u003Cp>Both plans give you Claude Code with the same tools, commands and workflows. What differs is the size of the allowance: Anthropic&#39;s help center describes a five-hour session limit and a weekly limit on both, with Max offering more room than Pro. Higher tiers also bring some model-specific allowances, and which models are included in the plan or billed to usage credits can differ by plan. The \u003Ccode>\u002Fmodel\u003C\u002Fcode> picker marks models that &quot;require usage credits&quot;, and extended context for some models depends on plan too. Look at the picker on your own account rather than a blog post (including this one) for the current state.\u003C\u002Fp>\n\u003Cp>Limits are shared across Claude chat and Claude Code. If you also use claude.ai heavily, that is part of the same pool, which is an argument for Max even if your Claude Code use alone is modest.\u003C\u002Fp>\n\u003Ch2 id=\"how-do-i-know-whether-i-need-max\">\u003Ca class=\"anchor\" href=\"#how-do-i-know-whether-i-need-max\" aria-label=\"Link to this section\">#\u003C\u002Fa>How do I know whether I need Max?\u003C\u002Fh2>\n\u003Cp>Use data, not feeling. For two weeks, note:\u003C\u002Fp>\n\u003Col>\n\u003Cli>\u003Cstrong>How often you hit a limit.\u003C\u002Fstrong> A session limit once a week that resets while you eat lunch is not a reason to upgrade. A limit every afternoon is.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>What was running when you hit it.\u003C\u002Fstrong> \u003Ccode>\u002Fusage\u003C\u002Fcode> breaks recent usage down by skills, subagents, plugins and MCP servers, and recommends changes. If one habit is behind a big share (a loop, an always-on subagent, a giant context), fix that first.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Whether you hit the weekly limit.\u003C\u002Fstrong> The 5-hour window is an annoyance; the weekly one stops you for days. Hitting it more than once in a while is the clearest upgrade signal.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>How many sessions you run at once.\u003C\u002Fstrong> Parallel sessions multiply usage. See \u003Ca href=\"\u002Fblog\u002Frun-multiple-claude-code-sessions\">running several Claude Code sessions\u003C\u002Fa>. Two or three on a big codebase is a different load from one.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>How much you wait.\u003C\u002Fstrong> Count the hours per week that Claude Code was stopped and you had work for it. Price that against the plan difference.\u003C\u002Fli>\n\u003C\u002Fol>\n\u003Ch2 id=\"can-i-cut-usage-before-upgrading\">\u003Ca class=\"anchor\" href=\"#can-i-cut-usage-before-upgrading\" aria-label=\"Link to this section\">#\u003C\u002Fa>Can I cut usage before upgrading?\u003C\u002Fh2>\n\u003Cp>Often, yes, and it is free. Most heavy usage comes from context size, not from the number of prompts: the whole conversation is re-sent with every request. Before you upgrade:\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Ccode>\u002Fclear\u003C\u002Fcode> between unrelated tasks and \u003Ccode>\u002Fcompact &lt;focus&gt;\u003C\u002Fcode> in long ones (\u003Ca href=\"\u002Fblog\u002Fclaude-code-context-window\">the context window, explained\u003C\u002Fa>).\u003C\u002Fli>\n\u003Cli>Lower the effort with \u003Ccode>\u002Feffort\u003C\u002Fcode> on simple tasks, and use a lighter model for routine edits and subagents.\u003C\u002Fli>\n\u003Cli>Keep CLAUDE.md short and move rarely needed instructions to skills.\u003C\u002Fli>\n\u003Cli>Use plan mode before large changes, so you do not pay for a wrong implementation.\u003C\u002Fli>\n\u003Cli>Stop loops and background agents you forgot about.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Cp>The full list is in \u003Ca href=\"\u002Fblog\u002Fclaude-code-usage-limits\">Claude Code usage limits\u003C\u002Fa>. If you do all of this and still hit the weekly limit, the upgrade is justified.\u003C\u002Fp>\n\u003Ch2 id=\"what-are-the-options-besides-changing-plan\">\u003Ca class=\"anchor\" href=\"#what-are-the-options-besides-changing-plan\" aria-label=\"Link to this section\">#\u003C\u002Fa>What are the options besides changing plan?\u003C\u002Fh2>\n\u003Cp>You are not choosing between only Pro and Max.\u003C\u002Fp>\n\u003Cul>\n\u003Cli>\u003Cstrong>Usage credits.\u003C\u002Fstrong> You can turn them on to keep working past a limit; they bill at API rates, with a monthly spend limit you set. \u003Ccode>\u002Fusage-credits\u003C\u002Fcode> opens the settings. Good for occasional overflow, expensive as a habit. If you overflow every week, a plan is likely cheaper. See \u003Ca href=\"\u002Fblog\u002Fclaude-code-token-costs\">what your usage would cost at API prices\u003C\u002Fa>.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Team or Enterprise seats.\u003C\u002Fstrong> If you work for a company, a seat has its own allowance, and admins can add usage credits and spend limits.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>A second account that is yours.\u003C\u002Fstrong> A personal plan plus a work seat, each with its own limits, is legitimate when they are for different purposes. See \u003Ca href=\"\u002Fblog\u002Fclaude-code-multiple-accounts\">using several Claude accounts with Claude Code\u003C\u002Fa>. Do not use extra accounts as a trick to dodge limits; read Anthropic&#39;s terms first.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>API billing (Console).\u003C\u002Fstrong> Pay per token, no windows. Predictable only if you watch it. For heavy interactive use, compare it with a plan using your own token numbers (see the token costs article above); it is also a different login mode from a subscription.\u003C\u002Fli>\n\u003Cli>\u003Cstrong>Waiting.\u003C\u002Fstrong> With automatic continue in recent Claude Code versions, a session stopped at a limit picks the task up after the reset if you keep it open.\u003C\u002Fli>\n\u003C\u002Ful>\n\u003Ch2 id=\"which-plan-for-which-kind-of-user\">\u003Ca class=\"anchor\" href=\"#which-plan-for-which-kind-of-user\" aria-label=\"Link to this section\">#\u003C\u002Fa>Which plan for which kind of user?\u003C\u002Fh2>\n\u003Cdiv class=\"table-wrap\">\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>If you...\u003C\u002Fth>\u003Cth>Lean toward\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>Use Claude Code a few hours a week, one session at a time\u003C\u002Ftd>\u003Ctd>Pro\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Mostly use chat and occasionally ask Claude Code for a fix\u003C\u002Ftd>\u003Ctd>Pro\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Hit the 5-hour limit most workdays\u003C\u002Ftd>\u003Ctd>Max\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Hit the weekly limit\u003C\u002Ftd>\u003Ctd>Max, after trimming context habits\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Run parallel sessions, worktrees or scheduled agents\u003C\u002Ftd>\u003Ctd>Max\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Work in a company with a Claude plan\u003C\u002Ftd>\u003Ctd>A team seat, plus your own plan for side work\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Need a pay-per-token bill for accounting reasons\u003C\u002Ftd>\u003Ctd>API (Console), knowingly\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003C\u002Fdiv>\n\u003Cp>Treat the first month on a higher plan as a test. You can compare how often you were blocked, and downgrade if the answer is never.\u003C\u002Fp>\n\u003Ch2 id=\"a-trap-to-avoid-the-api-key\">\u003Ca class=\"anchor\" href=\"#a-trap-to-avoid-the-api-key\" aria-label=\"Link to this section\">#\u003C\u002Fa>A trap to avoid: the API key\u003C\u002Fh2>\n\u003Cp>If \u003Ccode>ANTHROPIC_API_KEY\u003C\u002Fcode> is set in your shell, Claude Code uses it (once you approve it) and bills the API, not your plan. People pay for Max and then burn API credits without knowing. Run \u003Ccode>\u002Fstatus\u003C\u002Fcode> to see which credential is active, and \u003Ccode>unset ANTHROPIC_API_KEY\u003C\u002Fcode> if you want the subscription.\u003C\u002Fp>\n\u003Ch2 id=\"in-relais\">\u003Ca class=\"anchor\" href=\"#in-relais\" aria-label=\"Link to this section\">#\u003C\u002Fa>In Relais\u003C\u002Fh2>\n\u003Cp>Relais drives the Claude Code login already on your machine, with Pro, Max or a team seat; it has no plan of its own for Claude usage and needs no API key. Everything it runs counts toward your plan&#39;s limits, and your 5-hour usage shows under your account in the sidebar. Running several sessions and agents at once is exactly what makes a plan feel small, so Relais lets agents stop starting new work past a share of your plan, and carries a session over to your next account when one runs out. See \u003Ca href=\"\u002Fpricing\">pricing\u003C\u002Fa> for how Relais and your Claude plan are billed separately.\u003C\u002Fp>\n",[15,18,21,24,27,30,33],{"id":16,"text":17},"what-is-actually-different-between-pro-and-max-for-claude-code","What is actually different between Pro and Max for Claude Code?",{"id":19,"text":20},"how-do-i-know-whether-i-need-max","How do I know whether I need Max?",{"id":22,"text":23},"can-i-cut-usage-before-upgrading","Can I cut usage before upgrading?",{"id":25,"text":26},"what-are-the-options-besides-changing-plan","What are the options besides changing plan?",{"id":28,"text":29},"which-plan-for-which-kind-of-user","Which plan for which kind of user?",{"id":31,"text":32},"a-trap-to-avoid-the-api-key","A trap to avoid: the API key",{"id":34,"text":35},"in-relais","In Relais",[37,42,47],{"slug":38,"title":39,"description":40,"category":7,"categoryId":8,"keyword":41,"date":10,"minutes":11},"claude-code-usage-limits","Claude Code usage limits: the 5-hour window and weekly cap","How Claude Code's 5-hour window and weekly limit work, what makes you hit them, how to check usage, and what to do the moment Claude Code stops.","claude code usage limits",{"slug":43,"title":44,"description":45,"category":7,"categoryId":8,"keyword":46,"date":10,"minutes":11},"claude-code-context-window","Claude Code's context window: why it fills up and what to do","Why Claude Code's context window fills up, what loads before you type, and the commands (\u002Fcontext, \u002Fcompact, \u002Fclear) that stop you running out mid-task.","claude code context window",{"slug":48,"title":49,"description":50,"category":7,"categoryId":8,"keyword":51,"date":10,"minutes":52},"claude-code-multiple-accounts","How to use several Claude accounts with Claude Code","Keep a personal plan and a work seat signed in to Claude Code at once with CLAUDE_CONFIG_DIR, switch cleanly, and avoid the API key trap. Commands included.","claude code multiple accounts",4,1794931098788]