A click on ChatGPT Ads costs between $3 and $5 at entry, and a thousand impressions between $25 and $60. Every campaign needs at least $25 a day, though there has been no overall minimum spend since May 2026. These are ranges seen at launch. OpenAI publishes no benchmarks.
Updated 25 October 2026.
What CPC is observed on ChatGPT Ads?
Between $3 and $5 per click at entry with the Clicks objective, a range seen when an account starts and not a guaranteed price.
CPC is set through a maximum bid at ad group level: you name the ceiling you are willing to pay for a click. What you actually pay sits below it, because the auction is second price (more on that below).
The range leaves a lot out. Your CPC will depend on your sector, on how relevant your ads are and on the competition for the conversations you are targeting. Two accounts in two sectors have no reason to pay the same for a click, and a single average, however tidy, hides precisely that gap.
And the CPM?
Between $25 and $60 per thousand impressions, with the Views objective. The default maximum bid documented by OpenAI sits at the top of that range.
CPM makes sense when the aim is to be present in qualified conversations rather than to buy traffic. It also helps you learn: which intent families generate impressions, and at what pace, before you commit a per-click budget. As with CPC, the range recalibrates within the first days of delivery on your own data.
A minimum budget of $25 a day per campaign
With no overall minimum spend since self-serve opened in May 2026, the only floor left is the campaign’s: $25 per day.
The daily budget behaves like an average. One day’s spend may overshoot the amount you set and another’s may fall short, so that the period lands on the average, and three campaigns with three different objectives therefore need three times the daily minimum, or $75 a day in all.
Before May 2026, getting into Ads Manager came with a spend commitment. The ticket is now low. An account can test the channel with a single campaign at the minimum, and the real cost of entry moves elsewhere: it is the time needed to structure intents, ads and tracking, far more than the media.
Why is it cheaper than Google Ads today?
Because a few tens of thousands of advertisers share ChatGPT Ads, against several million on Google Ads: fewer bidders per conversation, so lower prices. The gap is temporary.
The mechanism is plain arithmetic. The price of a click at auction depends on how many advertisers want it. On Google Ads that number reflects years of competitive optimisation; on ChatGPT Ads it reflects a channel that opened to European advertisers on 31 August 2026.
The gap will close as advertisers arrive. Better to see it as a window. Those who learn while it is open keep a structural advantage once it shuts, because they already know which intents, which ads and which pages work.
Relevance as a price lever
The auction is second price and weighted by relevance, so a more relevant ad can win the placement with a lower bid than a competitor that is less well aligned.
OpenAI describes the mechanism as a relevance-weighted second-price auction. Each candidate ad gets a score combining its bid with its relevance to the current conversation, that score decides who is shown, and the winner pays what it took to beat the runner-up, never its own ceiling.
For the budget, this plays out in three ways. A precise title and a body that adds information raise the score and lower the price, and a landing page that carries the conversation forward has the same effect. An ad group that mixes two products does the reverse: a lower score, a higher price and fewer placements. The creative is managed like a cost line. The mechanism is covered in more detail in our article on how ChatGPT Ads works.
Five cost lines, only one of them visible in Ads Manager
Media, ad production, landing pages, tracking and, where relevant, agency fees. Only the first appears in Ads Manager.
| Cost line | What determines it |
|---|---|
| Media (CPC or CPM) | Chosen objective, maximum bid, ad relevance, number of advertisers on the same conversations |
| Ads | Number of intent families to cover, title and body variations, square visuals to produce |
| Landing pages | One page per intent rather than a generic homepage; accessibility to OpenAI’s crawlers |
| Tracking | OpenAI Pixel installation, server-side Conversions API, UTM convention, CRM reconciliation |
| Agency fees | Billing model: flat fee, percentage of media or a mix |
The native report covers only the first line, in six columns: impressions, clicks, spend, CTR, average CPC, average CPM. Cost per lead, per demo or per order is something you work out yourself, using tracking. The native report stops short of it.
Why are there no official benchmarks?
Because OpenAI states in its own documentation that it publishes no performance benchmarks, whether by advertiser, by industry or by campaign type.
The native report gives you your average CPC without saying whether that CPC is good for your sector. The platform publishes no “average e-commerce CPC” and no “average SaaS CPC”, and any sector range you read elsewhere comes from a player with its own accounts, or from one that is extrapolating.
With no published reference, the only workable comparison comes from accounts actually being run, sector by sector. That is why the Hiints Index exists: it gathers the CPC, CPM, CTR and cost per opportunity observed across our accounts. The two ranges quoted at the top of this article, $3 to $5 a click and $25 to $60 per thousand impressions, remain entry markers. We promise no CPC level, and the index is no substitute for working out your own cost per opportunity.