LLM API Cost Calculator
Estimate the USD cost of an LLM API workload from input/output token counts and daily call volume, using published per-model list prices — with per-call, per-day, per-month and per-year projections and a side-by-side model comparison.
Input
Request
Tokens sent to the model per call (prompt + context).
Tokens the model returns per call (completion).
How many requests you make each day. Used for daily and monthly projections.
Model & Pricing Mode
Output
Cost for Selected Model
| Metric | Value |
|---|---|
| No data yet | |
Side-by-side Model Comparison
| Model | Provider | In / Out ($/1M) | Per Call | Per Day | Per Month |
|---|---|---|---|---|---|
| No data yet | |||||
Pricing is based on published list prices per 1M tokens and may be outdated — confirm against the provider's current pricing page. Batch pricing applies the provider's standard 50% discount (OpenAI, Anthropic, Google); Meta / Llama hosted APIs typically have no batch tier.
Guides
Estimate what an LLM API workload will cost before you ship it. Enter the number of input tokens (your prompt plus context) and output tokens (the model's completion) for a single call, pick a model, and set how many calls you expect per day. The calculator returns the cost per call, per day, per month and per year, plus a side-by-side comparison across every model in the table so you can see instantly whether a cheaper model would do.
How it works
The cost of one call is:
cost = (inputTokens / 1,000,000) × inputPricePerMillion
+ (outputTokens / 1,000,000) × outputPricePerMillion
Daily cost multiplies that by your calls-per-day; monthly uses an average of 30.4375 days per month, and yearly multiplies the monthly figure by 12. Output tokens are almost always billed at a higher rate than input tokens, which is why a short prompt with a long completion can cost far more than the token counts alone suggest.
Real-time vs batch pricing
Many providers offer a batch tier at roughly 50% off for work that doesn't need an immediate response (OpenAI, Anthropic and Google all do). Switch the pricing mode to Batch to apply that discount to the models that support it. Meta / Llama models are priced here from common hosted providers and typically have no batch tier, so their price is unchanged in batch mode.
A note on prices
The built-in price table is a snapshot and may be out of date. LLM pricing changes frequently — providers cut prices, add models and retire old ones. Treat the figures here as a planning approximation and always confirm the current rate against the provider's official pricing page before committing to a budget. Prices are list prices in USD per 1 million tokens and don't reflect enterprise discounts, cached-input rates, or per-region differences.
Which token counts should I use?
Use the average tokens per call for a representative request. If you don't know your token counts yet, most providers publish a tokenizer or a token-count endpoint; as a rough rule of thumb, one token is about four characters or three-quarters of an English word.
Why is the output cost higher than the input cost?
Output (completion) tokens are generated one at a time and are more expensive to produce, so providers price them higher — often 3–5× the input rate. Reducing the length of the model's responses is usually the fastest way to cut cost.
Privacy
This tool runs entirely in your browser. Your token counts and settings are never sent to a server.