LLM API Cost Calculator
Estimate what a model call costs per request, per 1,000 calls, per 1,000,000 calls, and per month. Enter token counts directly or paste a prompt for an instant token estimate.
Anthropic Claude Opus 5: $5.00 / 1M input tokens • $25.00 / 1M output tokens
The math
input: 1,000 / 1,000,000 x $5.00 = $0.005000
output: 500 / 1,000,000 x $25.00 = $0.0125
per call = $0.005000 + $0.0125 = $0.0175
monthly = per call x 200 calls/day x 30 days = $105.00
How LLM API billing works
Almost every hosted model bills two separate meters: prompt (input) tokens and completion (output) tokens, each quoted per million tokens. Output is priced higher because the model runs one forward pass per generated token. To estimate a call, convert both token counts to millions, multiply by the matching rate, and add them. Multiply by your call volume for daily and monthly projections. Real invoices also include retries, multi-turn tool calls, system prompt overhead, and sometimes a discount for cached or batched input, so treat this as a planning envelope rather than an exact forecast. The single biggest lever is usually the output cap: halving max tokens on a chat endpoint often halves the bill. Trimming retrieved context is the next lever. Pair this with the Token Counter to size prompts and the tools hub for more utilities.