Claude Code Pricing and Rate Limits (2026): Plans, Caps, and How to Track Them

Claude Code pricing starts at $20/month on the Claude Pro plan and scales to $100–$200/month on Max plans, with usage governed by a rolling 5-hour session window plus a weekly cap introduced by Anthropic in 2025. Developers who run into limits mid-session should track usage proactively rather than guessing when the next reset lands.

• Pro plan: $20/month, shared Claude Code + chat usage, lowest rate-limit ceiling.
• Max 5x / Max 20x: $100/month and $200/month, multiplying the Pro usage ceiling by 5x or 20x.
• API (pay-per-token): No subscription cap, billed per input/output token by model tier (Opus, Sonnet, Haiku).

For people who want a live view of where they stand against these caps instead of waiting for a "limit reached" error, Usagebar is a macOS menu-bar app that reads your /usage data and shows the 5-hour window, weekly cap, and context usage in real time.

What does Claude Code cost?

Claude Code is bundled into Claude's consumer subscriptions rather than sold separately. You get access to it on the $20/month Pro plan, the $100/month Max 5x plan, or the $200/month Max 20x plan, and usage is shared with claude.ai chat sessions on the same account. Anthropic also offers Team and Enterprise tiers with seat-based pricing for organizations (Anthropic pricing page).

Plan	Price	Relative usage limit	Best for
Pro	$20/month	1x (baseline)	Light, occasional Claude Code use
Max 5x	$100/month	5x Pro	Daily coding work, single project
Max 20x	$200/month	20x Pro	Heavy, multi-project / team-adjacent use
API	Pay-per-token	No fixed cap, rate-limited by tier	CI pipelines, custom integrations, agents

If you're deciding between paying for a subscription or metering usage through the API instead, that trade-off is covered in more depth in does Claude Code usage affect Pro limits.

How do Claude Code rate limits work?

Claude Code rate limits run on two layers: a rolling 5-hour session window that caps how much you can use in a short burst, and a weekly cap that caps total usage across the billing cycle. Both limits reset on fixed schedules, and hitting either one blocks new prompts until it clears.

The 5-hour window starts counting from your first message in a session and resets exactly 5 hours later, not at a fixed clock time, per Anthropic's own documentation on using Claude Code with your Pro or Max plan. The weekly cap, added in August 2025 after reports of heavy automated usage, resets on a 7-day cycle tied to your subscription renewal date rather than a calendar week. For a deeper breakdown of this timing, see when does Claude Code usage reset and how often does Claude usage reset.

What happens when you hit the limit?

When you hit either the 5-hour or weekly cap, Claude Code stops accepting new prompts and shows a message telling you when the limit resets. You can keep working by switching to the API with pay-per-token billing, waiting out the reset window, or upgrading to a higher Max tier for more headroom.

There is no official way to force an early reset on a subscription plan. Some workarounds and their limitations are detailed in can you reset Claude usage limit early, and the specific stop-and-wait behavior inside the CLI is covered in Claude Code stop and wait for limit to reset.

How much does the API cost per token?

Claude's API bills per million tokens and varies by model: Opus-tier models cost the most per token, Sonnet-tier models sit in the middle, and Haiku-tier models are cheapest, with output tokens generally priced several times higher than input tokens. Exact per-model rates are listed on Anthropic's pricing page and change periodically as new model versions ship.

API usage doesn't share the 5-hour/weekly subscription caps — instead it's governed by requests-per-minute and tokens-per-minute rate limits tied to your usage tier, documented at docs.anthropic.com. This makes the API a common fallback once a subscription's weekly cap is exhausted, though at the cost of unpredictable monthly spend compared to a flat subscription fee.

How can you check your usage before hitting the cap?

You can check Claude Code usage three ways: running /usage inside the CLI, viewing the usage dashboard at claude.ai/settings/usage, or installing Usagebar for a persistent menu-bar view that updates without you having to ask. The CLI and web dashboard both require you to actively check; Usagebar surfaces the same data passively.

Usagebar tracks the 5-hour session window, the weekly cap, and context-window usage, with alerts at 50%, 75%, and 90% so you get warned before work gets interrupted mid-task. It's a one-time $9 purchase, not a subscription. Full setup steps are in how do I check my Claude Code usage limits.

For developers burning through caps faster than expected, pairing usage tracking with consumption habits matters too — see how to reduce Claude Code token usage for practical ways to cut context bloat.

Key takeaways

• Claude Code is bundled into Pro ($20/mo), Max 5x ($100/mo), and Max 20x ($200/mo) subscriptions, not sold as a standalone product.
• Two rate-limit layers apply: a rolling 5-hour window and a weekly cap, both resetting on a schedule tied to your account, not the calendar.
• Hitting a cap blocks new prompts until reset; the API (pay-per-token) is the main fallback for continuing work immediately.
• You can check usage via /usage, the claude.ai dashboard, or Usagebar's always-on menu-bar display.
• As of November 2026, Anthropic has not published fixed numeric caps (e.g. exact hours or token counts) for Pro/Max tiers — limits scale dynamically with model and demand.

If reset timing and rate-limit surprises are the main pain point, Get Usagebar for a live, always-visible read on your 5-hour window, weekly cap, and context usage — no more guessing when Claude Code will let you back in.

Sources

• Using Claude Code with your Pro or Max plan – support.claude.com
• Anthropic pricing page
• Anthropic developer docs