LLM Cost Calculator
Model a real API workload with input, cached, output, and reasoning tokens. Add request volume, Batch share, retries, context tiers, and a monthly budget.
01 / Input
Token workload
02 / Scenario
Monthly workload
Route comparison
API in One vs official pricing
Method
A workload estimate, not a token counter
Token rates are stored by model and price tier, in USD per million tokens. Cached reads, cache writes, output, and long-context rates stay separate so discounts are not applied to the wrong category.
Monthly projections include the selected request period, additional retry attempts, and the discounted share eligible for Batch. Usage imports are normalized to per-request averages before simulation.
FAQ
LLM cost calculator questions
How accurate is the text token estimate?
Text mode uses a browser-side heuristic for planning. For an invoice-grade result, import the provider usage object or enter tokenizer output manually.
Are uploaded files sent to a server?
No. Supported text files are read locally in the browser and are not uploaded by this calculator.
How are cache and Batch discounts applied?
Cached input uses the model's published cached-input rate. Batch affects only the selected share of the workload and only when that price catalog supports a Batch discount.
Which usage exports can I import?
The parser accepts JSON, JSONL, and CSV with common OpenAI, Anthropic, and Gemini token fields, including cached and reasoning token details.