- Set the workload
Enter monthly active users and any other volume assumptions that define the period or batch.
- Add usage assumptions
Complete the remaining usage fields, keeping token, duration, or size units consistent with their labels.
- Enter current rates
Use the current prices for the exact model, endpoint, quality, and processing tier you plan to use.
- Review the estimate
Read the main result and the supporting breakdown to see the largest cost or capacity driver.
- Test another scenario
Change one assumption at a time or select Reset to restore the default example values.
OpenAI Chatbot Calculator
The OpenAI Chatbot Calculator estimates monthly message volume, token usage, API cost, and peak request rate for a conversational application. It combines active users, conversations, messages, token lengths, and model prices.
This supports capacity and budget planning for support bots, internal assistants, and consumer chat products. The result is an estimate because real conversations vary in length and may include cached context, tools, retrieval, moderation, or retries.
Calculation inputs
users
conversations
messages
tokens
tokens
USD
USD
%
Result
—
Estimated result
Monthly messages—
Monthly tokens—
Average peak RPM—
Cost per active user—
Monthly messages = Active users × Conversations per user × Messages per conversation. Cost = Messages × [(Input tokens × input rate + Output tokens × output rate) ÷ 1,000,000].
The calculator applies the entered values directly and displays rounded results for planning.
What the result means
The main result is the estimated monthly text-model charge for chatbot messages.
Peak RPM is an averaged planning indicator, not a guarantee that rate-limit requirements will be met.
Given: Example assumptions
Calculation: With 20,000 users, four conversations each, and six messages per conversation, monthly volume is 480,000 messages. At 900 input and 250 output tokens with rates of $2.50 and $15 per million, cost per message is $0.006, producing an estimated monthly cost of $2,880.
Result: The displayed result follows the same formula and rounding logic.
Should one user turn count as one message?
Use one billable model response as a message. If a turn triggers several model calls, multiply accordingly.
How should conversation history be counted?
Include the average history sent with each request in input tokens, not only the newest user message.
Does the estimate include retrieval and tools?
No. Add embedding, file search, web search, code execution, storage, or other tool costs separately.
Why estimate peak traffic share?
It converts a portion of monthly volume into an average requests-per-minute planning figure for the busiest traffic period.
Can I compare two models?
Yes. Keep workload assumptions unchanged, enter each model’s rates in separate runs, and compare cost and context constraints.