Claude Pricing Explained: Free, Pro, Max, and Ultra Plans Compared
Claude's free, Pro, Max, and Ultra plans explained: weekly limits, session caps, API credits, and which tier actually fits your workflow.
What are Claude’s pricing tiers?
Anthropic sells Claude access across four tiers: a free plan, Pro at $20 a month, Max at $100 a month, and Ultra at $200 a month. Each step up multiplies your usage allowance rather than opening up a totally different product. Pro is the baseline paid tier, Max gives five times Pro’s weekly usage, and Ultra gives ten times Pro’s weekly usage. The free plan gets you Claude’s Sonnet and Haiku models but not Opus, and it won’t run Claude Code sessions the way paid plans do.
TL;DR
- Claude runs on two limits at once, a 5-hour session cap and a weekly cap, and both are shared across the Claude chat app and Claude Code, so a long coding session and a casual chat draw from the same pool.
- Ultra ($200/month) gives you 10 times Pro’s weekly usage, and verified Ultra accounts no longer have a 5-hour session limit at all.
- Max ($100/month) gives 5 times Pro’s weekly usage, which is the middle option for people running Claude Code regularly but not nonstop.
- Max and Ultra both include monthly API credits ($100 for Max, $200 for Ultra), but those credits are for building your own apps programmatically, not for extra usage inside the Claude Code or chat apps.
- The free plan includes Sonnet and Haiku, plus chat features like web search, memory, connectors, and artifacts, but it does not include Opus.
- Opus can eat up to half your weekly limit on Max and Ultra plans, while Pro users pay for Opus access through usage credits instead of a flat allowance.
- You can check exactly where you stand at any time by typing
/usagein Claude or clicking the usage circle in the bottom right of the desktop app.
Nothing gets lost.
How do the weekly and session limits actually work?
Claude enforces two overlapping limits. The first is a 5-hour session limit, a rolling window that caps how much you can do in a short burst. The second is a weekly limit, which caps your total usage over seven days. Both limits are shared across every surface you use: a Claude Code session building out a project, and a plain chat conversation, draw from the exact same bucket. There’s no separate allowance for “chat” versus “code.”
This matters because people often assume Claude Code is a separate product with its own quota. It isn’t. If you spend most of your weekly budget running long Claude Code sessions, you’ll have less room left for chat-based work, and vice versa.
Ultra is the one plan that changes this structure meaningfully: verified Ultra accounts no longer have a 5-hour session cap, so the constraint that matters is purely the weekly limit. For Pro and Max, both limits still apply.
What’s the actual difference between Pro, Max, and Ultra?
The difference is almost entirely about volume, not features. All three paid tiers get access to Claude Code, the chat app, and the model lineup (Sonnet, Haiku, and Opus, with usage of the top-tier Opus model metered differently depending on plan).
- Pro ($20/month): the entry paid tier. You get Claude Code access and Opus usage, but Opus is paid for through usage credits rather than a flat weekly allotment.
- Max ($100/month): five times Pro’s weekly usage. This is aimed at people who run Claude Code regularly, have multiple sessions going, or do heavier day to day work.
- Ultra ($200/month): ten times Pro’s weekly usage, plus the removal of the 5-hour session limit for verified accounts. This is built for people running many parallel sessions or long-running agentic tasks throughout the day.
On Max and Ultra, Anthropic’s most capable model (a top-tier reasoning model, referred to in Anthropic’s materials as built for demanding reasoning and long autonomous tasks) can consume up to half of your weekly limit if you lean on it heavily. That model is also significantly more expensive on a pay-per-use basis than Opus, roughly two and a half times the cost, which is part of why Anthropic gates it behind a portion of your weekly allowance rather than unlimited access.
What are the API credits for, and why don’t they help inside the app?
Max and Ultra subscriptions include monthly API credits: $100 worth for Max, $200 worth for Ultra. These credits are not extra fuel for the Claude Code or chat apps. They’re meant for the Anthropic API, which is what you’d use to build your own applications or automations on top of Claude, things that run independently of the Claude interface, such as a script or background service hitting the model programmatically.
| Vendor | Tier | Risk | ApprovalNEW |
|---|---|---|---|
| Northwind Supply | Tier 1 | Low | Approved |
| Acme Logistics | Tier 2 | Medium | Approve |
| Helios Data | Tier 1 | High | Pending |
| Granite Freight | Tier 3 | Low | Approved |
| Pinecrest Labs | Tier 2 | Low | Approve |
This is a common point of confusion. People sometimes expect that running low on their weekly chat/Code limit means they can dip into their API credits to keep going inside the app. That’s not how it works. The two pools are separate: one covers interactive usage inside Anthropic’s own apps, the other covers raw API calls for things you build yourself.
Is the free plan actually usable, or just a demo?
The free plan is more functional than many people expect. It includes Sonnet and Haiku (though not Opus), and as of recent updates it also includes web search, memory, connectors, and artifacts, the same artifact system that lets Claude generate documents, slides, and simple interactive tools directly inside a conversation. Docs, slides, and basic design generation are also available on free, not gated behind a paywall.
What the free plan does not give you is Claude Code in any serious capacity, or Opus-level reasoning for harder tasks. If your work involves multi-file coding projects, long agentic sessions, or tasks that benefit from Opus’s reasoning, you’ll hit the ceiling quickly. For lighter use, quick questions, drafting documents, simple design mockups, free is genuinely workable.
Is Max or Ultra worth it for heavier users?
Whether Max or Ultra make sense depends almost entirely on how much Claude Code or long-session work you do. If you’re running a handful of Claude Code sessions a week for writing, light automation, or document work, Pro may be sufficient, especially if you keep the model’s effort level set to medium instead of maxing it out unnecessarily (higher effort settings consume more of your limit per task without proportionally better results in many cases).
If you’re running many parallel sessions, doing extended agentic work where Claude operates across files for long stretches, or regularly pushing into Opus territory, Max’s five-times multiplier buys meaningful headroom. Ultra’s ten-times multiplier combined with no 5-hour cap is really built for people treating Claude as an always-on part of their workflow, running a dozen or more sessions across a day, rather than occasional use.
The API credits bundled into Max and Ultra are a secondary consideration. If you’re not building your own programmatic integrations, those credits go unused, which effectively makes the all-in price per month of “interactive usage” higher than the sticker price suggests.
Frequently Asked Questions
What’s the difference between the 5-hour limit and the weekly limit?
The 5-hour limit is a rolling short-term cap on usage, while the weekly limit tracks total usage over seven days. Both apply across chat and Claude Code combined. Verified Ultra accounts have the 5-hour limit removed entirely, leaving only the weekly cap.
Does the free plan include Claude Code?
The free plan does not provide the same Claude Code access that paid plans do. Claude Code, along with the full Opus model and Claude’s agent features, is positioned as a paid-plan capability, starting at Pro.
What’s the difference between usage limits and API credits?
Usage limits (5-hour and weekly) govern how much you can do inside Claude’s own apps, chat and Claude Code. API credits, included with Max and Ultra, are a separate monthly allowance for calling Anthropic’s API directly to build your own tools or automations. They don’t top up your in-app usage limit.
How do I check how much of my limit I’ve used?
Built like a system. Not vibe-coded.
Remy manages the project — every layer architected, not stitched together at the last second.
Type /usage inside a Claude conversation, or click the usage indicator (a small circle) in the bottom right corner of the desktop app. This shows where you stand on both the session and weekly limits.
Is Opus available on every plan?
Opus is not included on the free plan, which offers Sonnet and Haiku. On Pro, Opus usage is billed through usage credits rather than a flat allowance. On Max and Ultra, Opus draws from your weekly limit, and can consume up to half of it if used heavily.
