Claude Cost Calculator
Model Anthropic Claude API spend across Opus, Sonnet, and Haiku.
Pricing dataset updated Oct 12, 2026
Estimated monthly cost
$14.75
- Cost per request
- $0.0003
- Input tokens
- $6.00
- Output tokens
- $8.75
- Annualised
- $177.00
Estimates use public list pricing per 1M tokens. Batch, enterprise, and regional rates may differ.
How it works
Choose a Claude model based on results from your own tasks, then estimate the cost of that workload. A coding assistant may resend source files and tool results over several turns, so the total tokens across the session matter more than the first prompt alone.
Compare a routine request with a longer agent session using separate estimates. Model routing can reduce spend, but savings disappear if a cheaper model causes more retries or additional human review. This calculator estimates token charges, not task success.
Worked example
Reviewing a coding-session budget
At 20,000 calls averaging 8,000 input and 1,500 output tokens, monthly volume is 160 million input and 30 million output tokens. With illustrative rates of $3 input and $15 output per million, the estimate is $480 + $450 = $930. If a session uses four calls, count all four; 20,000 sessions would cost $3,720 under these same assumptions. This is a planning example, not a current price guarantee.
Provider references: Anthropic prompt caching