How much do ChatGPT Ads cost?

Price is not economics

There are three different questions hidden inside “What do ChatGPT Ads cost?” How does the platform bill? What price will this particular campaign achieve? And what acquisition cost can the business afford? Answering one does not answer the others.

As reviewed on 8 September 2026, OpenAI’s campaign documentation describes CPM, CPC and conversion-optimized CPC objectives. For oCPC, billing is per valid click, not per conversion. It also distinguishes daily budgets from campaign-total budgets. Source: Create campaigns.

That documentation does not establish a universal CPC benchmark for your offer. Check account-specific settings and the current currency minimums before launch. This guide deliberately avoids presenting a market-wide price range as a forecast.

Work backwards from value

Start with contribution economics and the conversion path. For a lead-generation business, an illustrative model is:

For example, suppose a business chooses an allowable acquisition cost of $1,200, expects 5% of leads to become customers, and assumes 4% of clicks become leads. The implied allowable cost is $60 per lead and $2.40 per click. These are invented planning inputs, not ChatGPT Ads benchmarks or forecasts.

The model is only as useful as its assumptions. A lead-to-customer rate from a mature referral channel may not transfer to paid traffic. A click-to-lead rate from a branded campaign may not transfer to a new audience. Use conservative, base and optimistic scenarios, then replace assumptions with cohort evidence.

For e-commerce, use net contribution rather than gross sales alone. Include returns, fulfillment and other variable costs that change what the business can afford to acquire a customer. Keep lifetime value assumptions separate from revenue already realized.

Budget for useful evidence

A test budget should be tied to the question the team wants to answer. A destination usability issue can sometimes be diagnosed with a small sample and qualitative evidence. A modest difference in acquisition efficiency may require far more observations.

Suppose a planning scenario assumes a 4% click-to-lead rate. At 100 clicks the expected count is only four leads, with substantial uncertainty. One additional lead changes the observed rate by a full percentage point. That is not a stable basis for declaring a winning channel.

Do not substitute an arbitrary spend threshold for experiment design. Define the comparison, likely baseline, minimum commercially meaningful change and review horizon. If the available budget cannot support the intended inference, narrow the question. You can learn whether a page is functioning before trying to estimate a reliable acquisition cost.

Read the full cost

Media is one part of the investment. Account for strategy, research, creative production, landing-page development, measurement setup and ongoing management. Separate one-time setup from recurring costs so a first-month comparison does not hide either burden.

Track staff capacity as well. A campaign that generates qualified consultations faster than a sales team can handle them may look efficient in an ad report while wasting demand in practice. The acceptable pace of growth belongs in the budget conversation.

OpenAI’s rules and controls can change. Recheck the campaign documentation when setting spend limits, rather than assuming a daily budget behaves like a hard daily ceiling. Keep the authorization and review process explicit within your team.

Set the stop rules

Write down what triggers intervention before the campaign starts: broken tracking, an unreachable destination, unexpectedly poor lead quality, or spend reaching the agreed test allocation without useful evidence.

A good review separates three outcomes: evidence to scale, evidence to revise, and insufficient evidence. The third is a valid result. It should lead to a better-designed test or a decision to pause, rather than a confident story built around a small sample.

For the operational sequence, see the campaign playbook. To discuss a test in the context of your acquisition economics, book a strategy call.