Published June 10, 2026
Claude Code is not free. The cheapest way in is the Claude Pro plan at $20/month, and it can also be billed pay-as-you-go per API token (Claude Opus 4.8 runs $5 input / $25 output per million tokens). This guide covers every way to pay for Claude Code in 2026: the subscription plans, the per-token API rates, the shared usage limits, and the June 15, 2026 Agent SDK credit change that puts a hard cap on automated (headless and CI) usage.
For free Claude access routes see Claude Free Credits 2026; for the new frontier model see Claude Fable 5.
Is Claude Code free?
No. There is no standalone free tier for Claude Code. Anthropic's Free plan ($0) includes chat and code generation in the Claude apps, but not Claude Code itself. Claude Code first appears on the Pro plan ($20/month).
Two real ways to pay:
- A Claude subscription (Pro, Max, Team, or Enterprise) - interactive Claude Code is included in the plan.
- An Anthropic API key - no subscription, billed per token at standard API rates through Claude Console.
How much does Claude Code cost? (subscription plans)
Plans that include Claude Code, verified 2026-06-10 against claude.com/pricing:
| Plan | Price | Claude Code |
|---|---|---|
| Free | $0 | Not included |
| Pro | $20/mo ($17/mo billed annually) | Included |
| Max (5x) | from $100/mo | Included, 5x Pro usage |
| Max (20x) | higher tier (commonly reported ~$200/mo, confirm on the live page) | Included, 20x Pro usage |
| Team - Standard seat | $25/seat/mo ($20 annual) | Included |
| Team - Premium seat | $125/seat/mo ($100 annual) | Included |
| Enterprise | $20/seat base + usage at API rates; contact sales | Included |
The "5x" and "20x" on Max mean 5x and 20x more usage than Pro. Note that on a subscription, Claude Code shares its usage allowance with Claude chat on the same account, it is one pool, not two.
Claude Code pay-as-you-go: API token pricing
If you skip the subscription and run Claude Code with an Anthropic API key, you pay per token at standard API rates. Current model rates, verified 2026-06-10 against Anthropic's pricing page:
| Model | Input /MTok | Output /MTok |
|---|---|---|
| Claude Haiku 4.5 | $1 | $5 |
| Claude Sonnet 4.6 | $3 | $15 |
| Claude Opus 4.8 | $5 | $25 |
| Claude Fable 5 | $10 | $50 |
You switch models inside Claude Code with /model, and yes, the new Fable 5 frontier model is selectable. Two cost levers to know:
- Prompt caching cuts repeat-context cost: cache reads are billed at 0.1x the input rate; the Batch API is 50% off.
- Tokenizer overhead: Opus 4.7 and newer use a tokenizer that can produce up to ~35% more tokens for the same text, so your effective per-page bill is higher than the sticker rate. Budget for it.
For most individuals, the $20 Pro subscription is cheaper than pay-as-you-go. API billing wins when usage is spiky, automated, or you want a hard per-token cost you can model.
What changed on June 15, 2026 (the Agent SDK credit)
This is the most important pricing change of the year for anyone automating Claude Code. Effective June 15, 2026 (official support article):
- Interactive Claude Code in the terminal or IDE, and normal chat, are unchanged, still included in your subscription.
- But Agent SDK usage,
claude -p(headless / non-interactive), Claude Code GitHub Actions, and third-party apps that authenticate through your subscription now draw from a separate monthly Agent SDK credit, billed at standard API rates:
| Plan | Monthly Agent SDK credit |
|---|---|
| Pro | $20 |
| Max 5x | $100 |
| Max 20x | $200 |
| Team Standard seat | $20 |
| Team Premium seat | $100 |
| Enterprise (usage / seat) | $20 / $200 |
The credit does not roll over. When it runs out, automated requests either spill to pay-as-you-go usage credits (if you have enabled them) or stop until the credit refreshes next month. Eligible users get an email to claim the credit before June 15.
Plain version: if you run Claude Code headless in CI, scripts, or GitHub Actions on a subscription, you now have a real automation cost cliff. Heavy automation users should either budget usage credits or move that workload to a straight API key with predictable per-token billing.
Usage limits and how billing works
- Shared limits, 5-hour reset. Claude Code and chat share the plan's usage limits, which reset on a rolling 5-hour window.
- May 6, 2026 increase: Anthropic doubled Claude Code's 5-hour rate limits for Pro, Max, Team, and seat-based Enterprise, and removed the peak-hours limit reduction for Pro and Max (announcement).
- Overflow: Pro and Max users can enable usage credits (Settings > Usage, or
/usage-creditsin the CLI) to keep working past plan limits at standard API rates, with a monthly spending cap. - API-key billing: every token is billed pay-as-you-go; Claude Code auto-creates a "Claude Code" workspace in Console so you can track spend. The dollar figure Claude Code shows in
/usageis for visibility, on a subscription it is not what you are billed.
Which Claude Code plan should you pick?
A practitioner's call:
- Trying it / light use: Pro at $20/month. Best value entry point.
- Daily heavy interactive coding: Max 5x ($100) or 20x for the larger usage pool and priority access.
- Teams: Team seats (Standard $25, Premium $125) for shared admin and billing.
- Automation-heavy (CI, headless, Agent SDK): after June 15 you are capped by the Agent SDK credit, so either enable usage credits or run that workload on a dedicated API key and pay per token, it is the only predictable option at scale.
Frequently asked questions
Is Claude Code free? No. There is no free Claude Code tier. The Free Claude plan includes chat and code generation but not Claude Code. The cheapest way to use Claude Code is the Pro plan at $20/month, or pay-as-you-go with an Anthropic API key (no subscription, billed per token).
How much does Claude Code cost per month? Pro is $20/month ($17 billed annually). Max starts at $100/month for 5x Pro usage, with a higher 20x tier (commonly reported around $200/month). Team seats are $25 (Standard) and $125 (Premium) per seat. Enterprise is contact-sales. All of these include Claude Code.
Can I use Claude Code without a subscription? Yes. Claude Code works with an Anthropic API key billed per token at standard rates: Haiku 4.5 $1/$5, Sonnet 4.6 $3/$15, Opus 4.8 $5/$25, Fable 5 $10/$50 per million input/output tokens. This is usually cheaper than a subscription only when usage is spiky or automated.
What is the June 15, 2026 Agent SDK credit change? From June 15, 2026, automated Claude Code usage (Agent SDK, claude -p headless mode, GitHub Actions, third-party apps via your subscription) draws from a separate monthly Agent SDK credit at API rates: $20 on Pro, $100 on Max 5x, $200 on Max 20x. Interactive terminal and IDE use is unchanged. The credit does not roll over.
Does Claude Code include Claude Opus or Fable 5? Claude Code lets you switch models with /model. On a subscription, available models and usage depend on your plan; on an API key you pay each model's token rate. Opus 4.8 and the new Fable 5 frontier model are both usable in Claude Code.
Which Claude Code plan is best? Pro ($20) for trying it and light use, Max (from $100) for daily heavy interactive coding, Team seats for teams, and a dedicated API key for automation-heavy or CI workloads where you want predictable per-token cost.
