Your AI agent re-reads its tool list on every step — that's 32% of the bill

A developer measured the token cost of resending an agent's full tool schema on every step of an agent loop, finding that 18 typical coding-agent tools consume 2,388 tokens per request and account for roughly 32% of a $0.150 ten-step run. Trimming the toolbox to the six tools a task actually needs cut the per-run cost to $0.121 and the monthly schema overhead from $1,433 to $563 at 1,000 runs a day.

You give your agent a toolbox: read a file, run the tests, search the web. The model picks a tool, your code runs it, the result goes back, the model picks the next one. It feels like the model knows its tools. It doesn't. It forgets them after every reply, and you pay to remind it. I counted what that reminder costs on a realistic toolbox. Short version: about a third of the bill. A few words first. A token is roughly ¾ of a word; APIs bill per million of them, one price for input what you send and a higher one for output what the model writes . Tool calling is the API feature where you describe functions and the model replies with "call read file with these arguments" instead of plain text. Each function is described with a JSON Schema — a JSON object listing the name, a description and the parameters with their types. Here is one of them, exactly as an OpenAI-compatible API expects it: {"type": "function", "function": { "name": "edit file", "description": "Replace an exact string in a file with a new string. The old string must appear exactly once in the file unless replace all is true. Preserve indentation exactly.", "parameters": {"type": "object", "properties": { "path": {"type": "string", "description": "Path to the file to edit."}, "old string": {"type": "string", "description": "Exact text to find. Must be unique in the file."}, "new string": {"type": "string", "description": "Text to replace it with."}, "replace all": {"type": "boolean", "description": "Replace every occurrence instead of requiring a unique match."}}, "required": "path", "old string", "new string" }}} The API is stateless: nothing is remembered between requests. So the whole tools array travels inside the input of every request, next to the system prompt and the history. An agent loop — the model calls a tool, gets the result, decides the next move — is one request per step. Ten steps, ten copies of the toolbox. I wrote 18 tools of the kind every coding agent ships with: read file , write file , edit file , list directory , search code , find files , run shell , run tests , three git tools, web search , fetch url , http request , query database , create issue , send message , ask user . Descriptions are ordinary, one or two sentences each. Then I counted them with tiktoken and o200k base , the tokenizer behind current GPT models: python import json, tiktoken pip install tiktoken enc = tiktoken.get encoding "o200k base" tools = json.load open "tools.json" your "tools" array total = len enc.encode json.dumps tools print f"{len tools } tools = {total} tokens per request" for t in tools: n = len enc.encode json.dumps t print f" {t 'function' 'name' :<16} {n}" Result: 2,388 tokens , 133 per tool on average. The smallest, git status with no parameters, is 54. The biggest, search code with five parameters, is 193. That's for one request — before the user has typed a word. Now a modest agent run: a 500-token system prompt, a 200-token task, 10 steps, each step adding about 780 tokens of history a tool call plus its result and producing 150 output tokens. Prices are GPT-5.6's official ones on 15.10.2026: $2.00 per million input, $12.00 per million output. | Toolbox | Tool tokens per run | Share of input | Cost per run | Schemas per month | |---|---|---|---|---| | All 18 tools | 23,880 | 36% | $0.150 | $1,433 | | 6 tools the task needs | 9,380 | 18% | $0.121 | $563 | | No tools baseline | 0 | 0% | $0.102 | $0 | 1,000 runs a day, 30 days, input price only. With the full toolbox, the schemas are $0.048 of a $0.150 run — 32% of the bill . Per request, they weigh more than three times the system prompt and the task combined. And it's not a one-off: at 1,000 runs a day that's 716 million tokens a month spent re-describing send message to a model fixing a typo. Ship the agent only the tools the job needs. A coding fix doesn't need create issue , send message or query database . Cutting to six relevant tools dropped the schema cost from $1,433 to $563 a month — $870 saved without touching a prompt. Second, trim the descriptions. I kept every tool but cut each description to its first sentence and removed per-parameter descriptions: 2,388 → 1,545 tokens, −35%. Test that the model still picks the right tool afterwards; descriptions are how it decides. Third, run the script above on your own tools array. It takes a minute and turns "it's just some JSON" into a number. The price per token is the other multiplier. We resell GPT-5.6 at $1.70 per million input against the official $2.00, so the same 716M tokens of schemas cost $1,218 instead of $1,433. That's 15% off a waste you shouldn't be paying for at all — trimming the toolbox comes first. My count is approximate. Providers don't send your JSON to the model verbatim: they rewrite it into an internal format and add their own instructions, so the real figure is a bit different and differs per vendor — check usage.prompt tokens on a real call with and without tools. Prompt caching changes the money a lot: the tool list sits at the start of every request, which is exactly what caching discounts, so with caching on, the 32% shrinks. I can't show how much on our traffic — our usage records don't split cached tokens out yet. And swapping toolboxes per step breaks that cache, so "send fewer tools" and "cache them" pull against each other. Count your tools before you count anything else. They're the part of the prompt you wrote once and pay for forever.