Anthropic Claude API Pricing: Complete Cost Guide 2026
Compare Claude model tiers, estimate monthly API usage, and review the pricing factors that affect a production workload.
Claude model tiers and their prices
Claude offers Haiku, Sonnet, and Opus models for workloads with different capability and cost requirements. Opus 5.5 and Sonnet 5.5 are newer options for complex work and coding tasks.
Consult the Anthropic pricing hub for the current input and output rates. Rates can vary by model, caching, context length, processing mode, and provider.
Monthly cost estimates for common usage levels
Estimate a monthly bill by multiplying input and output token volumes separately by the applicable rates per million tokens. For example, a workload with 10 million input tokens and 5 million output tokens costs 10 times the input rate plus 5 times the output rate.
Coding agents and document workflows can consume substantial input through repeated context and tool calls. Include retries and cached input when estimating your actual workload.
How Claude compares to GPT-5 and Gemini on price
Compare Claude Opus 5.5 with GPT-6.1 Sol and Gemini 4 Argon using the linked model comparisons. Check both input and output prices and the assumptions behind the estimate.
A lower token rate does not guarantee a lower cost per successful task. Include retries, context size, output length, and task quality in the comparison.
Choosing the right Claude tier for your product
Test a smaller model for classification, extraction, and routing, then measure its accuracy on representative examples.
Evaluate Sonnet or Opus on more demanding tasks and compare the cost per successful result. Confirm current prices and availability with Anthropic before budgeting.
Related pages
Estimate your own workload
Use the calculator to compare your expected API bill with a purchased or rented GPU setup.
Open calculator