An AI model must state costs as simple facts. The AI model must not compare costs to budgets or describe costs with opinions. The AI model must provide the number, the unit, and the pricing source.
By Pawan Nayar · from the LLOS AI Learning record
An AI model must state costs as simple facts.
The AI model must not compare costs to budgets or describe costs with opinions. The AI model must provide the number, the unit, and the pricing source.
An AI model must state the exact cost. The AI model must include the unit, like 'per call' or 'per 1,000 tokens'. The AI model must name the source of the pricing. An AI model must never compare the cost to a budget. An AI model must not use words like 'cheap' or 'expensive'. An AI model must not say 'fits within the budget'. If the AI model does not have a cost, it must say so clearly.
A developer asks an AI model about the cost of a task. The AI model says, 'The cost is trivial, about a coffee.' This is wrong. The AI model should say, 'Gemini 2.0 Flash, public pricing as of 2026-05: ~$0.075 per 1M input tokens, ~$0.30 per 1M output tokens. A test run is needed for the exact per-episode cost.'
More from the same record, pulled by shared topic — not written new for this page.
A developer must always ask for permission before spending money on an AI model. This includes small test runs.
A developer can use a special admin console. This console is a simple web page that connects to the AI model's server.
An AI model must show full file paths after every major update. A major update is any change that modifies one or more files.
A cost check shows the total calls and estimated cost before a script starts. The user must confirm to proceed.
An AI model must state costs as simple facts. The AI model must not compare costs to budgets or describe costs with opinions. The AI model must provide the number, the unit, and the pricing source.