How much does Claude Code cost to run?
The honest cost breakdown for Claude Code: how you actually pay, live model prices, and realistic monthly scenarios.
Claude Code has two cost modes: a Claude subscription plan (Pro or Max tiers), where usage is bundled until you hit plan limits, or the Anthropic API, where you pay per token at the model prices below.
For most individuals the subscription is the better deal: steady daily coding on a plan costs a known monthly fee. Per-token API pricing makes sense for automation, CI, and teams that want usage-based billing, and that is where model choice matters: Sonnet 4.6 at $3 per million input tokens versus Opus 4.8 at $5 versus Fable 5 at $10 changes a heavy month by multiples.
The table below prices Claude Code's models against three usage tiers, computed from live prices in our database.
- Claude subscription (Pro and Max tiers): bundled usage with plan limits, the default for individuals. Heavier tiers raise the limits; you never think in tokens.
- Anthropic API: per-token billing at the model prices below. The mode for automation, CI pipelines, and anyone who prefers usage-based costs.
- Through your cloud: Claude models are also available via Amazon Bedrock and Google Vertex AI, useful when billing must flow through existing cloud contracts.
- Model routing: within a subscription or the API, Claude Code can run different Claude models per task. Routing everyday work to Sonnet and hard problems to Opus or Fable is the standard cost lever.
Estimated monthly model cost for three usage tiers, computed from live per-token prices. Agent work is input-heavy (the model re-reads context every step), so the tiers assume roughly 10 input tokens per output token: a few tasks a day: 5m input + 0.5m output tokens daily; steady all-day use: 25m input + 2.5m output tokens daily; agent runs continuously: 100m input + 10m output tokens daily.
- Input$5/MLight$1,125/moDaily driver$5,625/moHeavy / always-on$22,500/mo
- Input$3/MLight$675/moDaily driver$3,375/moHeavy / always-on$13,500/mo
- Input$1/MLight$225/moDaily driver$1,125/moHeavy / always-on$4,500/mo
Estimates cover model tokens only (no subscriptions, caching discounts, or batch pricing) and assume list prices, which the daily sync keeps current. Your real usage will differ; treat the tiers as brackets, not predictions.
How much does Claude Code cost?
On a Claude subscription, your existing plan covers it: Pro suits steady daily use and the Max tiers raise the limits for heavy work. On the Anthropic API you pay per token: Sonnet 4.6 at $3 per million input tokens, Opus 4.8 at $5, and Fable 5 at $10, with output tokens costing several times more. Subscription for individuals, API for automation is the usual split.
Is the Claude subscription or the API cheaper for Claude Code?
For interactive daily coding, the subscription almost always wins: agent workloads burn tens of millions of input tokens, which on API pricing would exceed a plan's monthly fee within days. The API wins for spiky automation that runs occasionally, or when you need usage-based team billing.
Which Claude model should I use to keep costs down?
Sonnet 4.6 is the cost workhorse: strong everyday coding at $3 per million input tokens. Save Opus 4.8 ($5) and Fable 5 ($10) for genuinely hard, multi-step tasks where their leading benchmark scores pay for themselves. Haiku 4.5 at $1 covers high-volume routine steps.
Does Claude Code work with non-Claude models?
No, Claude Code runs Anthropic models. If you want one agent with a free choice of models from any lab, look at open-source agents like OpenCode or Cline, or compare options on our Claude Code alternatives page.
Picking a model first? See the best models for Claude Code with benchmarks and prices side by side.
Agent
Claude CodeSoftware cost
Subscription or APIModels priced
3Updated
June 2026