ClaudeStore/Docs
⚠ These docs are a work in progress. Some content may be inaccurate or temporarily AI-generated.

Claude API Pricing Explained

A clear breakdown of how Claude API pricing works — tokens, credits, models, and cost optimization tips.

How Claude API Pricing Works

Claude API pricing is based on tokens — the basic units of text that the model processes. Every request has two components: input tokens (what you send) and output tokens (what Claude generates). You pay for both.

On ClaudeStore, Credits are deducted from your balance based on actual token consumption per request.

What Are Tokens?

A token is roughly 4 characters of English text, or about ¾ of a word. For example:

  • "Hello, world!" ≈ 4 tokens
  • A 1,000-word article ≈ 1,300 tokens
  • A typical code file ≈ 500–2,000 tokens

Token Categories

CategoryWhat It IsRelative Cost
Input (uncached)Standard input tokens sent to the modelBase rate
OutputTokens generated by Claude in the response3–5× input rate
Cache WriteInput tokens written to prompt cache (first request)1.25× input rate
Cache ReadInput cached from a previous request0.1× input rate
ThinkingExtended thinking tokens (Opus/Sonnet)Same as output rate
Cache read tokens are up to 90% cheaper than uncached input. Use prompt caching for repeated system prompts to dramatically reduce costs.

Per-Model Pricing

Different Claude models have different price points. Anthropic's official list rates are per million tokens (MTok), while ClaudeStore bills at about 30% below those official rates:

ModelInput / 1M tokensOutput / 1M tokensBest For
Claude Opus 4.6$3.50$17.50Complex reasoning, research, architecture
Claude Sonnet 4.6$2.10$10.50Coding, general tasks, best value
Claude Haiku 4.5$0.70$3.50Fast responses, classification, autocomplete

Cost Calculation Example

Here's how to estimate the cost of a typical request using Claude Sonnet 4.6 at ClaudeStore's public rate ($2.10/MTok input, $10.50/MTok output):

Example: 2,000 input + 500 output tokenstext
Input: 2,000 tokens × $2.10 / 1,000,000 = $0.0042
Output: 500 tokens × $10.50 / 1,000,000 = $0.00525
─────────────────────────────────────────────
Total: $0.00945 per request

At this rate, $5 (1,000 credits) gets you ~529 similar requests.
That same request would deduct about 1.89 credits.

Tips to Reduce Costs

1. Choose the Right Model

Don't use Opus for tasks that Sonnet or Haiku can handle. Sonnet is 5× cheaper than Opus and handles most coding and general tasks excellently.

2. Use Prompt Caching

If you send the same system prompt repeatedly, enable prompt caching. After the first request, cached tokens cost only 10% of the standard input rate.

3. Optimize Prompt Length

Remove unnecessary context from your prompts. Shorter prompts = fewer input tokens = lower cost.

4. Set Per-Key Limits

Use ClaudeStore's per-key credit limits to prevent unexpected spending, especially on development and testing keys.

5. Monitor Usage

Check the ClaudeStore dashboard regularly to understand your consumption patterns and optimize accordingly.

Is ClaudeStore the Cheapest Claude API Option?

ClaudeStore offers one of the most affordable ways to access the Claude API. Unlike Anthropic's direct API, there's no minimum spend, no waitlist, and no KYC. You can start with as little as $5 in credits.

Because ClaudeStore uses the same underlying API, you get identical model quality at competitive rates. Combined with prompt caching (90% savings on repeated prompts), it's a cost-effective choice for developers and teams looking for a cheap Claude API key without compromises.

For a detailed comparison of ClaudeStore vs Anthropic direct API pricing, see our Claude API vs Direct guide. You can also check our Cheapest Claude API breakdown.

Compare with Other Providers

For a detailed comparison of ClaudeStore vs Anthropic direct API, see our Claude API vs Direct guide.

Ready to start?

Get API access to all Claude models in under 2 minutes.

View Plans