Claude API Pricing Calculator 2026
Best value: strong performance at 5x lower cost than Opus · $3.00 / 1M input · $15.00 / 1M output · 200K context
Show prompt caching prices
Cache writes cost 25% more. Cache reads cost 90% less.
Estimated monthly cost
$105.00
Per API call
$0.01
Monthly (10,000 calls)
$105.00
Annual (estimated)
$1,260.00
Input (1,000 × 10,000 calls)$30.00
Output (500 × 10,000 calls)$75.00
Output tokens cost 2.5x more than input for this model. Shortening responses with max_tokens is the fastest way to cut costs.
Switching to Claude Haiku 4.5 for the same workload would save approximately $77.00 per month.
Your annual spend projection is $1,260.00. Review whether all tasks require this model tier, or route simpler calls to Haiku 4.5.
Claude model pricing reference (2026)
Compare Claude vs OpenAI costs| Model | Context | Input / 1M | Output / 1M |
|---|---|---|---|
| Claude Opus 4.6 | 200K | $15.00 | $75.00 |
| Claude Opus 4.7 | 200K | $15.00 | $75.00 |
| Claude Sonnet 4.6best value· selected | 200K | $3.00 | $15.00 |
| Claude Haiku 4.5 | 200K | $0.80 | $4.00 |
Prices are per million tokens. Check anthropic.com/pricing for the latest rates.
About Claude API pricing calculator
The Claude API Pricing Calculator estimates your monthly Anthropic API spend based on model choice, average tokens per call, and expected call volume. All 2026 Claude models share a 200,000 token context window, and choosing between Haiku, Sonnet, and Opus based on task complexity can significantly reduce costs without sacrificing quality.
Anthropic prices input and output tokens separately, with output tokens costing more in every model tier. This calculator accounts for both, and includes a prompt caching toggle to show cached versus uncached rates. Enter your average prompt length, expected response length, and monthly call volume to get a reliable cost projection before committing to production scale.
How it works
Features
2026 model pricing
Opus 4.6, Opus 4.7, Sonnet 4.6, Haiku 4.5 with current rates.
Prompt caching toggle
See cache write and cache read rates alongside standard pricing.
Input/output breakdown
Separate costs for prompt and completion tokens.
Monthly projection
Scale from per-call to monthly and annual estimates.
Cost insights
Tips to reduce your bill based on your current model and volume.
Free and private
No account, no data stored.
How much does the Claude API cost in 2026?
Claude Sonnet 4.6 costs $3.00 per 1 million input tokens and $15.00 per 1 million output tokens as of June 2026. Claude Haiku 4.5 is the most affordable option at $0.80 input and $4.00 output per 1 million tokens.
All Claude 2026 models share the same 200,000 token context window, which is larger than GPT-4o at 128,000 tokens. Claude Opus 4.6 and Opus 4.7 are priced at $15.00 input and $75.00 output per 1 million tokens, designed for the most demanding reasoning and analysis tasks where accuracy is the priority.
For a side-by-side comparison with OpenAI models, use the OpenAI API cost calculator. To estimate costs before writing your prompt, the AI token counter shows token counts for Claude and other models instantly.
What is prompt caching and how much does it save?
Prompt caching lets Claude store repeated context between API calls so you are not re-billed for the same input on every request. Cache reads cost 90% less than standard input tokens, making it highly effective for applications with long system prompts or repeated document context.
Cache writes cost 25% more than the base input price, but savings start from the very next request. For a Sonnet 4.6 call with a 50,000-token system prompt, the standard input cost is $0.15 per call. With caching, subsequent calls cost $0.015 for that same prompt, a 90% reduction. The toggle above shows the exact cached versus uncached cost for your token count.
For multi-model cost comparisons including caching scenarios, the LLM cost comparison tool shows costs across Claude, OpenAI, and Gemini side by side.
Which Claude model is the best value in 2026?
Claude Sonnet 4.6 is the best value for most production applications in 2026, offering strong capability at $3.00 per 1 million input tokens. It handles coding, writing, analysis, and multi-step reasoning well at a fraction of Opus pricing.
Claude Haiku 4.5 is the right choice when volume is high and tasks are straightforward: classification, extraction, summarization, or customer support routing. At $0.80 per 1 million input tokens, it costs about 73% less than Sonnet 4.6. Reserve Opus 4.6 or Opus 4.7 for tasks where output accuracy is critical and cost is secondary.
Frequently asked questions about Claude API pricing
How much does the Claude API cost in 2026?+
Claude Sonnet 4.6 costs $3.00 per 1 million input tokens and $15.00 per 1 million output tokens. Claude Haiku 4.5 is the most affordable at $0.80 input and $4.00 output per 1 million tokens.
What is the difference between Claude Sonnet and Claude Haiku?+
Claude Sonnet 4.6 is the best value for most production tasks, balancing speed and quality at $3.00 per 1M input tokens. Claude Haiku 4.5 is faster and cheaper at $0.80 per 1M input tokens, suited for high-volume simple tasks.
What is prompt caching and how much does it save?+
Prompt caching lets Claude store repeated context between API calls. Cache reads cost 90% less than regular input tokens. Cache writes cost 25% more than regular input, but the savings on subsequent reads make it worthwhile for long system prompts.
How does Claude API pricing compare to OpenAI GPT-4o?+
Claude Sonnet 4.6 costs $3.00 per 1M input tokens versus GPT-4o at $2.50. Output costs $15.00 for Claude versus $10.00 for GPT-4o. Claude offers a larger 200K context window compared to GPT-4o at 128K.
What is the context window for Claude models in 2026?+
All Claude 2026 models support a 200,000 token context window. This is larger than GPT-4o at 128,000 tokens and makes Claude well suited for processing long documents, books, or codebases in a single request.
Is Claude Opus 4.6 worth the higher price?+
Claude Opus 4.6 costs $15.00 per 1M input tokens and $75.00 per 1M output tokens. It is designed for the most demanding tasks requiring deep reasoning, complex analysis, and high accuracy. For most production use cases, Sonnet 4.6 offers better value.
How do I calculate my monthly Claude API bill?+
Multiply your average input tokens per request by daily request volume and by 30 days for input cost. Do the same for output tokens. Add both totals and apply the per-million rate for your chosen model. This calculator does it automatically.
Does Anthropic offer free Claude API credits?+
New Anthropic accounts may receive small trial credits to test the API. Ongoing production use requires a paid account. Claude.ai Pro provides access to Claude models in the chat interface but does not include API credits.
Which Claude model should I use for my application?+
Use Claude Haiku 4.5 for high-volume simple tasks like classification and summarization. Use Claude Sonnet 4.6 for most production workloads requiring good quality at reasonable cost. Reserve Opus 4.6 for complex reasoning tasks where accuracy matters most.
How does Claude handle very long documents with its 200K context window?+
Claude can process up to 200,000 tokens in a single API call, equivalent to roughly 150,000 words or a full-length book. For repeated processing of the same long document, use prompt caching to reduce costs by up to 90% on the cached portion.