Anthropic's Claude has four pricing paths: a free tier, two consumer subscriptions (Pro and Max), a Team plan, and the pay-per-token API. This guide breaks down what you actually get for each, what the rate limits are, and which one fits your use case.
| Plan | Price | Best for |
|---|---|---|
| Free | $0 | Trying Claude, light daily use |
| Pro | $20 / month | Regular use, students, writers, light coding |
| Max (5x) | $100 / month | Heavy coders using Claude Code daily |
| Max (20x) | $200 / month | Power users at the rate-limit ceiling |
| Team | $30 / user / month | Small teams sharing usage |
| API (pay per token) | $1–$25 per 1M tokens | Apps integrating Claude, variable volume |
All consumer plans are billed monthly with an annual discount of roughly 15-17 percent. Enterprise pricing is custom.
Yes. The free tier at claude.ai gives you access to Claude Sonnet with limited daily messages. Anthropic does not publish a fixed message cap — the limit depends on prompt length, current load, and whether you're on web or mobile. In practice, free users typically get 20-40 messages every few hours before the system asks them to upgrade or wait.
The free tier does not include:
Claude Pro is the standard subscription. It costs $20 / month (or about $17/month billed annually). What you get:
Yes, if any of these apply: you hit the free-tier limits during a normal workday, you regularly need Opus-level reasoning, you work with documents larger than 30K words, or you want to share projects with the same context across many chats. If you only use Claude for occasional questions, the free tier is enough.
Claude Max comes in two tiers built for users who hit the Pro rate limits. Anthropic launched it specifically because Claude Code power users were burning through the Pro plan in hours.
If you use Claude Code professionally, the math usually breaks down like this: $200/month of Max covers what would cost $400-800/month on the pay-per-token API for an active developer.
The Anthropic API is pay-as-you-go, charged separately for input and output tokens. Current prices:
| Model | Input / 1M tokens | Output / 1M tokens | Context window |
|---|---|---|---|
| Claude Opus 4.7 | $5.00 | $25.00 | 1,000,000 |
| Claude Sonnet 4.6 | $3.00 | $15.00 | 1,000,000 |
| Claude Haiku 4.5 | $1.00 | $5.00 | 200,000 |
Claude Opus 4.7 is $5 per million input tokens and $25 per million output. For a typical Claude Code session with 50K input and 20K output, that's $0.75 for a single conversation. For a heavy day of 50 such sessions, around $37 in API cost — which is why Max ($200/mo) is the better option for that workload.
If you exhaust your Pro or Max message budget within a usage window, Anthropic does not currently allow paid overage on the consumer plans — you wait until the window resets (every 5 hours on most plans). For continuous use, the API is the only way to "pay extra" beyond your tier.
Anthropic's prompt caching feature lets you mark a prefix of your prompt as cached. When you reuse the same prefix within 5 minutes (or 1 hour on the extended TTL), input tokens for that prefix cost 10 percent of the normal price. This matters enormously for RAG pipelines, agents that reuse system prompts, and Claude Code (which caches the working-directory context).