How Much Do Claude Usage Credits Cost? API Pricing vs Pro/Max Plans Explained
In this article
Claude usage credits are prepaid, pay-as-you-go balances you buy on the Anthropic Console and spend per token through the API — not a flat subscription fee. As of November 2026, Claude Sonnet 4.5 costs $3 per million input tokens and $15 per million output tokens, while Claude Opus 4.5 runs $5/$25 per million tokens. That's different from the Claude.ai Pro plan ($20/month flat) or Max plan ($100–$200/month flat), which bundle a capped amount of usage into one price instead of metering every token. For most Claude Code users, a flat plan plus a tool like Usagebar to watch the limit is cheaper and simpler than buying API credits.
- Claude Sonnet 4.5 API: $3 / $15 per million input/output tokens
- Claude Opus 4.5 API: $5 / $25 per million input/output tokens
- Claude Pro subscription: $20/month flat (no per-token billing, usage capped on a rolling 5-hour window + weekly limit)
How much do Claude usage credits cost?
Claude usage credits are billed per million tokens based on which model you call, not a flat fee. As of November 2026, Sonnet 4.5 costs $3 per million input tokens and $15 per million output tokens; Opus 4.5 costs $5 input / $25 output per million tokens; Haiku models are priced lowest, typically under $1 per million input tokens. Credits are purchased in advance on the Anthropic Console and drawn down as you use the API, per the official Anthropic pricing page.
| Model | Input (per 1M tokens) | Output (per 1M tokens) |
|---|---|---|
| Claude Opus 4.5 | $5.00 | $25.00 |
| Claude Sonnet 4.5 | $3.00 | $15.00 |
| Claude Haiku 4.5 | ~$0.80 | ~$4.00 |
Prices move over time as Anthropic ships new model versions, so check the current rate card before budgeting a large credit purchase.
What's the difference between usage credits and a Claude subscription?
Usage credits are metered API billing: you pay per token, with no cap other than your wallet. A subscription (Pro, Max, Team) is a flat monthly fee that bundles a fixed usage allowance into Claude.ai and Claude Code, reset on a rolling schedule rather than drawn down token-by-token.
Credits make sense for programmatic API use where volume is unpredictable. Subscriptions make sense for day-to-day Claude Code or claude.ai use where you want predictable cost. We covered this trade-off in more depth in Claude usage credits vs subscription.
How much do Claude Pro and Max plans cost?
Claude Pro costs $20/month ($17/month billed annually) and includes Claude Code access with a shared usage pool. Max plans cost $100/month (5x the Pro limit) or $200/month (20x the Pro limit), per Anthropic's pricing page. Usage on every tier resets on a rolling 5-hour window, with an additional weekly cap layered on top.
| Plan | Price | Relative usage |
|---|---|---|
| Free | $0 | Baseline, limited daily messages |
| Pro | $20/mo | 1x |
| Max 5x | $100/mo | 5x Pro |
| Max 20x | $200/mo | 20x Pro |
For exact token allowances per tier, see how many tokens does Claude Pro give and how much usage do you get with Claude Pro.
When do you need to buy extra usage credits for Claude Code?
You need API usage credits when you hit your Pro or Max limit mid-session and want to keep working without waiting for the reset, or when you're running Claude Code in a context (CI, scripted agents, heavy batch jobs) that a subscription tier doesn't cover. Anthropic lets Claude Code fall back to pay-per-token API billing once a subscription limit is exhausted, per the official Claude Code + Pro/Max guide.
This is also where cost surprises happen: Opus-heavy sessions on metered credits can burn through $10–$30 in an afternoon of heavy refactoring, versus a flat $20/month Pro fee. Check whether Claude Code usage affects your Pro limits before assuming credits are cheaper.
How can you avoid buying credits you don't need?
Watch your 5-hour and weekly usage before you hit the wall, so you only reach for paid credits when it's actually necessary. Usagebar is a $9 one-time macOS menu-bar app that tracks Claude Code's 5-hour window, weekly cap, and context usage directly, with alerts at 50%, 75%, and 90% so you're never guessing whether the next prompt will throw a limit error.
That's cheaper than a single heavy day of metered Opus credits, and it means you can plan around the free reset window instead of defaulting to pay-per-token billing. For command-line checks instead, see how to check Claude Code usage limits or the built-in /usage command.
Get Usagebar to track usage before you need to buy a single credit.
Key takeaways
- Usage credits = metered, pay-per-token API billing (Sonnet 4.5: $3/$15 per million tokens; Opus 4.5: $5/$25 per million tokens as of November 2026).
- Subscriptions = flat monthly fee (Pro $20, Max $100/$200) with a rolling 5-hour + weekly usage cap, not per-token billing.
- Claude Code can fall back to credits once a subscription limit is hit, but Opus-heavy credit usage can cost more than a month of Pro in a single session.
- Tracking usage with reset timing and a menu-bar tool like Usagebar avoids unnecessary credit purchases.
Sources
Find Out Before You’re Locked Out, Not After
Checking your usage means stopping what you’re doing to go look. Usagebar keeps your 5-hour and weekly limits in the menu bar while you work, and alerts you at 50%, 75%, and 90% — so a lockout never lands mid-task.
Get Usagebar$9 — one-time, lifetime updates