ChatGPT visual ads are the newest format in OpenAI’s advertising business, and the first place they will appear is the moment people wait for a picture. On 5 October 2026 OpenAI announced an image-led ad unit that will sit beside, or just beneath, the images users ask ChatGPT to generate. Testing starts later in October, in the United States only, with an initial group of advertisers.
The announcement came with two other pieces that matter just as much to anyone who buys media: a long list of new measurement partners, and brand suitability pilots with DoubleVerify and Integral Ad Science. Put together, they show OpenAI trying to make ChatGPT a mainstream ad channel, with the reporting and safety assurances that big advertisers expect from Google or Meta. The company says ChatGPT now reaches 1.2 billion people a week.
This guide explains what OpenAI launched, how ChatGPT visual ads will look and behave, why image generation is the first placement, what the early performance figures do and do not prove, and what marketers and businesses in the UK should do while the format is still a US test. We have covered the wider debate about advertising inside AI chatbots before; this article is about the new visual format itself.
Table of contents
- What OpenAI Announced About ChatGPT Visual Ads
- How ChatGPT Visual Ads Will Look Next to an Image
- Why OpenAI Is Expanding ChatGPT Ads Now
- The Measurement Push Behind ChatGPT Visual Ads
- The Early ChatGPT Visual Ads Numbers, Read With Care
- Brand Safety and Privacy Around ChatGPT Visual Ads
- ChatGPT Visual Ads in the UK and Europe
- What ChatGPT Visual Ads Mean for Marketers
- What ChatGPT Visual Ads Mean for Users and Businesses
- ChatGPT Visual Ads FAQ
- References
What OpenAI Announced About ChatGPT Visual Ads
OpenAI published the news in a post titled “Building advertising for the way people use AI”, with a longer companion piece on its ads blog. The headline item is the new format, but the post is really three announcements in one.
The format in OpenAI’s own words
OpenAI describes the new unit as “a new visual ad format” that “helps people imagine how products and services could fit into their lives”. The ads use images showing “product inspiration, product usage, or the experiences they make possible”. For advertisers, the company says, “this creates new opportunities to tell their story through images.”
That is a deliberate shift. Until now, ChatGPT ads have been mostly text-led sponsored units placed under answers. ChatGPT visual ads put a picture at the centre of the ad, which is how most consumer brands already prefer to sell.
Where the first ChatGPT visual ads will appear
The test has one placement. “Initially, we’ll test this new ad format during image generation in ChatGPT,” OpenAI wrote. “Ads will be clearly labeled, and remain separate from the image being created.” The company repeated its standing line that “advertising does not influence the answers ChatGPT provides.”
Who will see them, and when
“Testing will begin later this month in the US with an initial group of advertisers,” the post says. OpenAI has not named those advertisers, and it has not said when ChatGPT visual ads will reach other countries. TechCrunch, which reported the launch, noted that the ads will eventually let brands reach ChatGPT’s full weekly audience “once they roll out globally”.
Here is what is confirmed and what is still open.
| Question | What OpenAI has said | Status |
|---|---|---|
| First placement | During image generation in ChatGPT | Confirmed |
| Start date | Later in October 2026 | Confirmed, no exact day |
| Market | United States only | Confirmed |
| Advertisers | An initial test group | Not named |
| Labelling | Clearly labelled, separate from the image | Confirmed |
| Pricing model | Not stated for the visual unit | Unknown |
| UK and EU launch | Not stated | Unknown |
How ChatGPT Visual Ads Will Look Next to an Image
OpenAI has not shipped the format yet, so the best guide to the experience is the set of mockups in its announcement. Mashable described them in detail.
Two layouts, one fictional grocer
According to Mashable’s write-up, the mockups show two layouts, both for a made-up brand called Heirloom Grocery. In the first, a single banner with a headline sits above a short product description and a “Learn more” button. In the second, a row of scrollable photos replaces the banner. Both carry an “Ad” label.
The carousel version matters for advertisers, because it lets one unit show several products or several uses of the same product. That is closer to a shopping ad than to a classic display banner.
The wait becomes the ad slot
The detail most coverage skipped is where the unit sits. In both mockups, the ad appears beneath a progress indicator for an image ChatGPT is still generating. In other words, the first ChatGPT visual ads fill the seconds a user spends watching a picture being made.
That is a clever slot. The user is already looking at the screen, already thinking visually and not yet reading a finished answer. It is also why OpenAI can claim the ad stays “separate from the image being created”: the picture and the ad occupy different parts of the interface, and the ad does not alter what the model draws.
Why image generation comes first
Image generation is a natural first home for ChatGPT visual ads for three reasons. People who generate images are often planning something concrete, such as a room, an outfit, a garden or a party, which is commercial intent in its purest form. The request itself reveals the subject without OpenAI needing to profile the user. And the surrounding interface is already visual, so an image ad looks like content rather than an interruption.
The obvious risk is the reverse of that last point. An ad that looks too much like the generated picture could blur the line OpenAI says it is protecting, which is why the labelling and placement rules carry so much weight.
Why OpenAI Is Expanding ChatGPT Ads Now
ChatGPT visual ads arrive eight months into a fast-growing ad business. The context explains both the timing and the push on measurement.
From a premium test to a mass channel
OpenAI began testing ads in the US in February 2026. Our earlier coverage of ChatGPT ad personalisation traced how the price fell as supply grew: a launch CPM of about $60, a range of roughly $25 to $35 by April, and some reported rates as low as $15. Minimum commitments of $200,000 to $250,000 fell to about $50,000 and were removed on 5 May 2026.
Supply has since grown sharply. The UK went live in June, a wider set of markets followed in August, and The Next Web reported on 31 August that ChatGPT Ads had passed a $1 billion annualised revenue run rate in under 200 days, with self-service buying live across Europe, India, the Middle East and North Africa.
The reported price path shows how quickly the channel has moved from scarce to scaled.
Bars are scaled to the $60 launch price: $35 is 58% of it, $25 is 42% and $15 is 25%. A falling CPM is good news for buyers, but it also means OpenAI needs new, more valuable inventory. A visual unit that brands can fill with product photography is exactly that.
Free and Go users carry the ads
Ads run only on ChatGPT’s Free and Go plans. Plus, Pro, Business, Enterprise and Education remain ad-free. TechCrunch put the trade-off bluntly: visual ads “could prove disruptive to the user experience”, and “perhaps that’s partly the point, as the ads could push more casual users to pay for a subscription in order to avoid them.”
That is the same two-sided model that powers YouTube and Spotify, and it is why we argued in the ugly economics of consumer AI that free chatbot use has to be paid for somehow. ChatGPT visual ads are one more lever: better revenue from free users, and a gentle nudge towards a paid plan.
Competition for the free audience
TechCrunch also linked the timing to Meta’s Muse, an agentic AI app that is free to everyone and a growing rival to ChatGPT. A free competitor raises the cost of making the free tier worse. OpenAI’s answer is to monetise free use more efficiently rather than restrict it, and richer formats such as ChatGPT visual ads earn more per user than plain text units.
The Measurement Push Behind ChatGPT Visual Ads
The less flashy half of the announcement may matter more to budgets. Big advertisers move money to channels they can measure, and OpenAI has spent the year building that plumbing.
Signals: pixel, Conversions API and data connectors
ChatGPT Ads already support a tracking pixel and a Conversions API (CAPI) for purchases, leads, installs and sign-ups. OpenAI has now added integrations with Hightouch, Tealium and LiveRamp, which “make it easier for advertisers to send conversion data from their existing systems to ChatGPT Ads”. An Event Quality Score helps advertisers check how good those signals are.
Attribution, incrementality and brand lift
OpenAI now lists ten attribution partners and three full-funnel measurement firms, and it is exploring geo-based incrementality tests. On the brand side, its ads blog says it is “in early-stage testing with Kantar and Cint to measure the impact of ChatGPT Ads on brand resonance metrics.”
| Job | Partners named by OpenAI | What it answers |
|---|---|---|
| Conversion data | Hightouch, Tealium, LiveRamp | Did the sale or lead reach ChatGPT Ads? |
| Attribution | AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge, Tenjin | Which ad gets credit for the outcome? |
| Full-funnel | Fospha, Measured, INCRMNTAL | How does ChatGPT fit the wider media mix? |
| Incrementality | Haus, Measured, WorkMagic | Would the sale have happened anyway? |
| Brand lift | Kantar, Cint (early testing) | Did the ad change how people see the brand? |
| Brand suitability | DoubleVerify, Integral Ad Science (pilots) | Are ads appearing in safe contexts? |
Measuring beyond the click
OpenAI argues that clicks undercount ChatGPT’s effect, because “clicking an ad may not be their immediate next step.” Advertisers can now pick a click-through attribution window and add a one-day view-through window. The company says 52.7% of eligible one-day view-through conversions in an early analysis happened within an hour of the matched ad impression.
That view-through argument will matter even more for ChatGPT visual ads, because an image ad is designed to be seen and remembered, not necessarily clicked while a picture is still loading.
The Early ChatGPT Visual Ads Numbers, Read With Care
OpenAI published four partner-reported results to support its case. None of them comes from the visual format, which has not launched, but they are the figures advertisers will hear in every sales pitch for ChatGPT visual ads.
The four figures OpenAI chose
DV Rockerbox measured WeightWatchers’ attributed cost per acquisition on ChatGPT Ads at 15.3% below its blended paid search benchmark. WorkMagic reported statistically significant lift for wellness brand Dose, with 67% of incremental purchases coming from net-new customers. Triple Whale found that 93% of Portland Leather’s visitors from ChatGPT Ads were new. And OpenAI’s own analysis gave the 52.7% view-through figure above.
Each bar below is a figure from OpenAI’s two posts, not our calculation.
Why one advertiser is not a benchmark
These are case studies, picked by the seller, from three named brands. They are useful as proof that measurement works, but they are not averages. A 15.3% lower CPA for one advertiser says nothing about the median advertiser, and a high share of new customers is normal for any new channel whose audience has not yet been saturated by retargeting.
The incrementality result is the strongest of the set, because geo-lift tests compare regions with and without ads rather than trusting attribution. OpenAI’s blog adds that WorkMagic’s study estimated 2.3 times more incremental orders for Dose than last-click attribution captured. Even so, it is one brand.
What to ask before trusting the figures
Before moving budget into ChatGPT visual ads on the strength of these numbers, ask the questions any media buyer would ask of a new channel. What was the test period? What share of spend did ChatGPT represent? Was the comparison benchmark paid search alone, or blended with brand terms that convert cheaply anyway? And how would the same view-through window flatter other display channels?
Brand Safety and Privacy Around ChatGPT Visual Ads
An ad that appears inside a personal conversation carries more risk than one beside a news article. OpenAI’s announcement spends real space on that problem.
Placement guardrails and Negative Phrases
OpenAI says its placement guardrails “assess whether a conversation is appropriate for advertising, helping keep ads out of emotionally vulnerable, sensitive, or otherwise unsuitable contexts.” Automated review, human oversight and ongoing monitoring sit behind that. For narrower needs, “Negative Phrases are now available to qualifying advertisers”, letting a brand exclude topics specific to its own policies.
Independent pilots without chat access
The brand suitability pilots with DoubleVerify and Integral Ad Science let those firms “assess how our brand safety standards are applied in a controlled testing environment, without accessing private user conversations.” IAS chief executive Lidiane Jones said in the announcement that marketers need protections that are “durable, consistent, and built to operate at scale.”
The design is a trade-off. Advertisers get outside verification, users keep their chats private, and neither side gets the full log-level transparency that verification firms provide on the open web. As with any new channel, ad fraud and impersonation are a cybersecurity concern too, so brands should treat the pilot results as a starting point rather than a guarantee.
What users control
OpenAI’s ads principles promise that “ChatGPT’s answers stay independent, conversations stay private, and users remain in control of their experience.” In practice, that means a setting for ad personalisation and the option to pay for an ad-free plan. Users who want no ChatGPT visual ads at all have one guaranteed route: upgrade to Plus or above.
ChatGPT Visual Ads in the UK and Europe
For a UK reader the most important fact is simple: ChatGPT visual ads are a US-only test for now. But British and European users already see ChatGPT’s text-led ads, and the rules here will shape any later rollout.
US-only, for now
OpenAI made the UK its first European ads market in June 2026, and TNW reported that ads have been running in 31 European countries since late August. The visual unit is not part of that yet. When it arrives, it will join a consent regime that is stricter than the US one.
The UK consent rule
In the UK and EU, OpenAI personalises ads only for users who explicitly opt in, relying on consent rather than legitimate interest as its legal basis. Users who decline still see ads, chosen from the current conversation, approximate location and time of day. Image generation suits that contextual model well: the prompt itself describes the subject, so ChatGPT visual ads could be relevant without any profile at all.
The EU’s search engine designation
On 31 August the European Commission designated ChatGPT a very large online search engine under the Digital Services Act, on the basis of 159.1 million monthly EU users. The DSA threshold is 45 million.
ChatGPT’s EU audience is 3.5 times the threshold (159.1 divided by 45). The designation brings a duty to publish a public ad repository covering a year of ads, including the advertiser, the period shown, the targeting parameters and the reach. Any ChatGPT visual ads served in the EU would land in that repository, which makes Europe the one place where outsiders will be able to study the format in detail.
What ChatGPT Visual Ads Mean for Marketers
If you sell consumer products, especially anything people picture before they buy, this format is worth preparing for now, even if your budget is in the UK.
Creative that shows the product in use
OpenAI’s own description points to three kinds of image: product inspiration, product usage and the experience a service enables. That is lifestyle photography, not packshots. Brands with strong in-context imagery for homes, fashion, food, travel and events are best placed for ChatGPT visual ads. The carousel layout in the mockups suggests a set of four to six images will serve better than a single hero shot.
Get your signals in order first
The measurement partners only help if conversion data flows. Before any test, install the ChatGPT Ads pixel or Conversions API, check the Event Quality Score, and make sure your attribution tool is one of the supported partners. Teams that already pipe events through a customer data platform can use the Hightouch, Tealium or LiveRamp connectors rather than building a new feed.
Plan a test, not a launch
Treat the first months of ChatGPT visual ads as an experiment with a control. A geo-split test, the method OpenAI itself is promoting through Haus, Measured and WorkMagic, is the most reliable way to see whether the channel adds sales rather than claiming credit for them. Our marketing services team can help design that kind of test.
| Preparation step | Why it matters | Owner |
|---|---|---|
| Build a lifestyle image library | The format sells use and experience, not packshots | Brand and creative |
| Install pixel or CAPI | No conversion data, no optimisation | Analytics or engineering |
| Check attribution partner support | Ten partners are supported today | Performance marketing |
| Write Negative Phrases | Keeps the brand away from unsuitable prompts | Brand safety |
| Design a geo-split test | Proves incremental sales, not just credit | Measurement lead |
| Agree a view-through policy | One-day view-through can inflate results | Finance and marketing |
Think about organic visibility too
Paid placement is only half of AI discovery. Brands also want to be named in ChatGPT’s unpaid answers, which depends on clear product information across the web. That discipline, often called AI optimisation, is covered by our AIO services, and it pays off whether or not you buy ChatGPT visual ads.
What ChatGPT Visual Ads Mean for Users and Businesses
The format affects the people on the other side of the screen as well: the users who generate images, and the organisations whose staff use ChatGPT at work.
For people generating images
Expect an “Ad” label beneath the progress indicator while a picture renders, on Free and Go accounts in the US first. The image you asked for should be unaffected. If an ad ever appears inside the generated picture rather than beside it, that would break OpenAI’s own stated rule, and it is worth reporting.
For organisations with staff on free accounts
ChatGPT visual ads are another reason to decide which ChatGPT tier your staff should use. Business and Enterprise plans carry no ads at all. Staff on personal free accounts will see ads, and their prompts help select them. If image generation is part of your team’s work, an approved paid tier keeps both the ads and the data question off the table. The same thinking applies to any generative AI tool with a free, ad-supported version.
The bigger picture for AI interfaces
ChatGPT visual ads show where AI interfaces are heading: the waiting time and the visual canvas inside an assistant are becoming inventory. Every company running a free assistant faces the same pressure to fund it. The open questions are whether labelling stays clear as formats get richer, and whether independent verification can work without seeing the conversations it is meant to protect.
ChatGPT Visual Ads FAQ
When do ChatGPT visual ads start?
OpenAI says testing begins later in October 2026 in the United States, with an initial group of advertisers. It has not given an exact date.
Where do ChatGPT visual ads appear?
During image generation. OpenAI’s mockups show the ad beneath the progress indicator for an image ChatGPT is still generating, labelled “Ad” and separate from the picture.
Will ChatGPT visual ads appear in the UK?
Not yet. The test is US-only. ChatGPT’s existing ads already run in the UK, where personalised ads need explicit consent.
Do ChatGPT visual ads change the image I asked for?
OpenAI says ads remain separate from the image being created and do not influence ChatGPT’s answers.
Which ChatGPT plans show ads?
Free and Go. Plus, Pro, Business, Enterprise and Education are ad-free.
How can advertisers measure ChatGPT visual ads?
Through the ChatGPT Ads pixel or Conversions API, ten attribution partners, three full-funnel measurement firms, geo-based incrementality tests with Haus, Measured and WorkMagic, and early brand-lift testing with Kantar and Cint.
How do businesses sign up?
OpenAI says businesses can sign up at ads.openai.com. Access to the visual test is limited to the initial advertiser group for now.
References
OpenAI: Building advertising for the way people use AI
OpenAI Ads: More ways to measure ChatGPT Ads
TechCrunch: OpenAI launches visual ads that appear alongside image generation results
Mashable: ChatGPT has new ad formats. Here’s what they look like
BleepingComputer: OpenAI will show visual ads in ChatGPT while you generate images
The Next Web: OpenAI is putting ads inside ChatGPT’s image generation
The Next Web: ChatGPT Ads pass a $1bn run rate as the EU designates ChatGPT
Digiday: ChatGPT ads land in the UK as OpenAI outlines EU privacy rules
European Commission: Digital Services Act questions and answers
