Grok Bot cost per task | Agent Receipts

What Grok Bot tasks cost

First round has not run

Tokens, pass or fail, and share of the weekly usage pool for 40 fixed everyday tasks. The first measurement round has not run yet, so this page has no results. We will not show estimates in the meantime.

Cost per task (empty until the first round)
Task	Median tokens	Min–max	Share of weekly pool	On-demand $	Passed	Valid runs
No results yet. The table fills in after the first round, with a receipt per task.

What each result will include

RECEIPT · TASK Txxtask set 2.1

Median tokens (last 3 valid runs)—

Min–max tokens—

Share of weekly pool—

Passed checklist—

Plan · model · version—

Measured on—

Raw usage export—

Tokens are the sum of all Grok Bot usage events in the window of one run.
Share of weekly pool is tokens divided by the weekly pool of the plan, recorded at the start of the week.
On-demand $ is only filled in when a run actually used paid on-demand usage. Grok Bot publishes no price per credit, so we do not convert tokens to dollars.
Fewer than 3 valid runs is marked provisional; older than 30 days is marked outdated.

All 40 prompts, inputs and checklists are public on the tasks page. How a run counts is on the method page.