Does Claude Code cost money on a Pro plan?
On Pro or Max, Claude Code is included in your flat subscription, so there is no per-use charge — but there are rate limits. Tracking 'cost' there means tracking how much of your allowance you are using.
Guide · Cost4 min read
How to track what Claude Code is actually costing you
Tracking Claude Code cost is not one task — it depends on how you pay. On a Pro or Max plan your cost is fixed and what matters is usage against your limits. On an API key you pay per token, so model choice drives the bill. Here is how to track each, and fold it into your whole AI spend.
Quick answer
How you track Claude Code cost depends on your billing. On a Pro or Max subscription the cost is flat, so 'tracking cost' means watching how much of your 5-hour and weekly allowance you are using. On an API key you pay per token, so you track usage by model (Opus costs more than Sonnet). Tokens 4 Breakfast reads local session token counts, shows usage against your plan limits, and estimates per-model spend — all from the menu bar, no API key required.
Step by step
The full walkthrough
Each step stands on its own — skip to the one that matches where you are.
Identify which billing you are on
Claude Code runs two ways: a Pro or Max subscription (a flat monthly fee with rate limits) or an Anthropic API key (pay per token). 'Cost' means something different for each, so start by knowing which one you use — or whether you use both.
On a subscription, track usage, not dollars
If you are on Pro or Max, your dollar cost is fixed. What actually matters is how much of your included usage you are burning, because hitting the 5-hour or weekly limit is the real constraint. Watch the limit meter, not an invoice.
On an API key, track tokens by model
If you pay per token, model choice dominates the bill: Opus is far more expensive than Sonnet, and cache reads are cheap. Tokens 4 Breakfast breaks your Claude Code tokens out by model so you can see exactly what is driving spend and switch models where it makes sense.
Fold Claude Code into your whole AI bill
Claude Code is rarely your only AI tool. Seeing it next to OpenAI, Cursor, Copilot, and your flat-rate subscriptions in one total is what stops surprises. Tokens 4 Breakfast keeps all of it in a single menu bar view instead of a dashboard per tool.
Set a budget and a forecast
Turn on a per-provider budget and a month-end forecast so drift is visible early. A budget alert at, say, 80 percent of where you want to land catches a runaway week before the month closes, not after.
Pro tips
- If you are on a subscription, do not chase dollar figures — usage against your limit is the number that can actually stop your work.
- Switching a routine task from Opus to Sonnet often cuts its token cost dramatically with little quality loss.
- A month-end forecast is more useful than today's total — it tells you where you are heading, not just where you are.
FAQ
Common questions
Short, direct answers to the things people ask most about this.
How do I see Claude Code cost by model?
Tokens 4 Breakfast reads your local Claude Code session logs and breaks tokens out by model, so you can see how much Opus versus Sonnet usage is contributing to your spend.
Can I track Claude Code cost alongside other AI tools?
Yes. Tokens 4 Breakfast is built to show Claude Code next to OpenAI, Cursor, Copilot, Gemini, and your subscriptions in one combined total.