Cook with AI Enrollment closed
AI & LLMs Runs locally

AI Agent Cost Visualizer

See how multi-step agent runs get expensive as context grows each step — per-step cost table with monthly totals.

ELI5

It shows why an AI agent can cost more every time it takes another step.

A little more detail

What this tool does

Agents often send the growing conversation, tool results, and prior work back to the model on every step. This calculator shows how that repeated context changes the cost of one run and a month of runs.

Use it to
  • Estimate the real cost of a multi-step agent workflow
  • See when growing context becomes the main expense
  • Test savings from fewer steps, shorter tool results, or cheaper models
Interactive workspace Results update as you type
Assumptions
Totals

Per run

Per day

Per month (30d)

Input tokens / run

Cost per step

Input billed = initial context + (step − 1) × tokens added. Each row is one LLM call in the agent loop.

Step Input tokens Output tokens Step cost Cumulative Relative

Pricing snapshot from July 2026 vendor pages. No prompt caching modeled — real agents with cache hits pay less on repeated prefix tokens.

Agent loops are sneaky expensive. Every step sends the full conversation plus tool results back to the model. If step 1 bills 4k input tokens and each step adds 1.5k of tool output, step 8 bills ~14.5k input — and you pay for all of it again.

That is why agent cost scales worse than linear: more steps mean longer context, and longer context means every subsequent step costs more. The fix is not always a cheaper model — it is summarizing tool results, trimming history, and caching stable prefixes.

This calculator uses simple arithmetic, not your real trace. Use it to sanity-check whether a 12-step ReAct loop at 100 runs/day is a rounding error or a line item.