AI Calculator Pro
Value

AI Chatbot Cost Calculator

Estimate the monthly cost of running an AI chatbot at scale.

Quick answer

For the “Small support bot” scenario, $201.15/month — $0.0335 per session · $0.00559 per message. Enter your own numbers below to recompute instantly; the full step-by-step math is shown under the worked example.

Results update automatically as you type.

Result
$1,503.75/month
$0.0501 per session · $0.00627 per message
Per message
$0.00627
Per session
$0.0501
Input cost / session
$0.0301
Output cost / session
$0.0200
Per day
$50.13
Per month(1,000 users/day)
$1,503.75

Estimate what it costs to run an AI chatbot. Because conversation history grows with every turn, cost per message rises through a session. Set your users, messages, and how you manage history to see a realistic monthly total.

How this is calculated

For each message we cost the input (system prompt + conversation history so far + the user message) plus the assistant's output, using your model's per-token rates. Because most chatbots resend the history on every turn, input tokens grow message by message — so the cost of a session is more than one message × the number of messages.

History handling changes the curve: 'full history' resends everything (most expensive), a 'sliding window' keeps only the last N turns, and 'summarized' compresses older turns. Session cost is summed across the messages, then multiplied by users per day × days per month for the monthly total.

Is this a good result? What to do next

Compare monthly cost to what each user is worth. For a free consumer bot the per-session cost is your true unit cost; for a paid product, cost per session should be a small fraction of revenue per user. If the monthly number surprises you, it's almost always the history mode or the response length driving it.

Typical planning ranges

Cost per session (mini model, windowed history)
~$0.002–$0.02
Cost per session (mid-tier model, full history)
~$0.03–$0.20
Realistic AI support-bot deflection
~20–50% of routine tickets

Ranges are typical planning figures to sanity-check your result, not authoritative benchmarks. Your numbers will vary with use case, volume, and vendor.

How to improve this number

  • Switch from full history to a sliding window or summarized history.
  • Cap assistant response length and tighten the system prompt.
  • Cache the system prompt so its tokens aren't re-billed every turn.
  • Route simple messages to a cheaper model; escalate only when needed.

Common mistakes

  • Costing one message and multiplying — this ignores history growth within a session.
  • Underestimating messages per session; real users go back and forth more than expected.
  • Assuming every daily user starts a full session (many bounce after one message).

When to use a different approach

For a single API call rather than a conversation, use the LLM API cost calculator. For agents that take tool-using steps, use the AI agent cost calculator.

Worked examples

Real scenarios you can reproduce by loading these presets in the calculator above:

Small support bot

Result
$201.15/month
$0.0335 per session · $0.00559 per message
Per message
$0.00559
Per session
$0.0335
Input cost / session
$0.0185
Output cost / session
$0.0150
Per day
$6.71
Per month(200 users/day)
$201.15
  • history mode: window (last 6 turns)
  • total input tokens/session: 7,410
  • monthly = session cost x 200 users x 30 days

Growth (10k users)

Result
$15,037.50/month
$0.0501 per session · $0.00627 per message
Per message
$0.00627
Per session
$0.0501
Input cost / session
$0.0301
Output cost / session
$0.0200
Per day
$501.25
Per month(10,000 users/day)
$15,037.50
  • history mode: window (last 6 turns)
  • total input tokens/session: 12,050
  • monthly = session cost x 10,000 users x 30 days

Enterprise (100k users)

Result
$200,175.00/month
$0.0667 per session · $0.00667 per message
Per message
$0.00667
Per session
$0.0667
Input cost / session
$0.0417
Output cost / session
$0.0250
Per day
$6,672.50
Per month(100,000 users/day)
$200,175.00
  • history mode: window (last 6 turns)
  • total input tokens/session: 16,690
  • monthly = session cost x 100,000 users x 30 days

Sources & references

Frequently asked questions

Why does chatbot cost grow within a session?+

Most chatbots resend the conversation so far on every turn. That means later messages carry more input tokens, so the per-message cost climbs unless you trim history.

How do I cut chatbot cost?+

Use a sliding window or summarized history, enable prompt caching for the system prompt, cap response length, and route simple messages to a cheaper model.

Does streaming change the cost?+

No. Streaming only changes how tokens are delivered; you still pay for the same input and output tokens.

Related calculators

API Cost
LLM API Cost Calculator

Estimate the cost of any LLM API call per request and per month.

Prompt Engineering
Context Window Planner

Check whether your prompt, history and context fit a model's window.

Agents
AI Agent Cost Calculator

Estimate the cost of multi-step AI agents with tool calls.

Answering a real question?

This calculator powers these problem-solving guides: