LLM API Cost Calculator
Enter current rates from your provider. The default rates below are illustrative examples, not live prices or quotes for a specific model.
Estimated total monthly operating cost
$6.00
$6.00
$0.20
$0.002
$72.00
This is an estimate, not a bill. Calculations run in your browser; no API key is requested or sent.
How to use the calculator
- Measure a representative request. Include system instructions, the user prompt, conversation history, retrieved context, and tool output sent back to the model.
- Enter billed token counts. Use tokens, not words. If requests vary widely, model a typical case and a high-usage case separately.
- Copy current rates from official sources. Pricing changes by model, context length, service tier, batch mode, region, and modality. Check OpenAI API pricing, Anthropic API pricing, or Gemini API pricing.
- Model caching only when applicable. Enter the cached-input share and cache-read rate from your provider’s billing definitions. Cache-write charges are not calculated separately.
- Add other costs and review time. Include hosting, storage, embeddings, search, per-call tools, or other charges not already covered by token rates. Add human review time when people must verify outputs.
Worked example: 3,000 requests per month
At 2,000 input tokens and 500 output tokens per request, 100 requests per day for 30 days uses an estimated 6 million input tokens and 1.5 million output tokens. With illustrative rates of $0.50 per million input tokens and $2.00 per million output tokens, input costs $3.00 and output costs $3.00, for $6.00 per month before caching, tools, infrastructure, taxes, or human review. These rates are examples only, not a quote for a named model.
What the estimate includes—and excludes
The calculator estimates text input and output, optional cached-input reads, additional monthly charges you enter, and review labor. It does not fetch live prices or connect to a provider account. It does not automatically include cache writes, image/audio/video billing, batch discounts, regional surcharges, minimum charges, retries, or every tool’s separate fee. Add known extra charges manually and compare the estimate with actual usage reports before budgeting.
Run a pilot on representative tasks, record billed usage and success rates, include failed or repeated calls, and model average and peak demand. A lower token price does not always mean a lower cost per successful task if a cheaper model needs more retries or human correction.
Related Aiverio resources
- Compare AI models by pricing, coding, and agent workflows.
- Explore AI computer-use tools and practical trade-offs.
- Read the Open WebUI self-hosting guide when evaluating locally hosted models.
- Browse the Aiverio AI tools directory.
Frequently asked questions
Does this estimate a ChatGPT, Claude, or Gemini consumer subscription?
No. It estimates usage-based API costs. Consumer subscriptions, enterprise contracts, credits, and third-party hosting plans have different billing rules.
How accurate is the result?
It is only as accurate as the token estimates and rates you enter. Provider invoices and usage dashboards are the source of truth. Measure a representative sample and allow for prompt variation, output variation, retries, and traffic spikes.
Does the calculator send my data anywhere?
No API key is required. Calculations run in the page’s browser using the values you enter. Do not paste private prompts, customer data, or secrets into this page.
