ChatGPT Will Test Image Ads During Image Generation This Month, and Advertisers' Inventory Complaint Stays
A.I. / news
ChatGPT Will Test Image Ads During Image Generation This Month, and Advertisers' Inventory Complaint Stays
OpenAI added a visual ad format and more measurement partners on October 5. The supply problem a media buyer cites is one ad slot and about one-sixth of Google's daily volume.
OpenAI said on Monday that it will begin testing a visual ad format in ChatGPT later this month in the US, with labelled image ads shown while a user generates an image. The announcement, dated October 5, also widens the measurement partners available to advertisers, according to BleepingComputer and Digiday.
OpenAI's own page blocked automated retrieval, so the details below are taken from those two outlets. Neither says which ChatGPT tiers will see the ads.
What the new format is
The test starts with an initial group of advertisers. OpenAI said the ads will show "product inspiration, how a product is used, or experiences associated with a service," and that they will be "clearly labeled, and remain separate from the image being created," as BleepingComputer quotes it.
Until now the format has been text. Digiday describes it as an advertiser name, logo, headline, description and landing page. OpenAI began selling ads in February 2026, and BleepingComputer puts ChatGPT's reach at 1.2 billion people a week.
Who measures the results
Two advertiser complaints are addressed in the update, according to Digiday: measurement that stops at last click, and a single format. The company is adding conversion tracking, attribution and brand-suitability partners.
| Function | Partners named |
|---|---|
| Conversion tracking | Hightouch, Tealium, LiveRamp |
| Attribution | AppsFlyer, Adjust, Branch, Triple Whale, Kochava |
| Geo-based incrementality tests | Haus, Measured, WorkMagic |
| Brand suitability pilots | DoubleVerify, Integral Ad Science |
The first three rows come from BleepingComputer and Digiday together; incrementality testing is reported by Digiday alone. OpenAI said the brand-suitability partners will not see private user conversations, and that safeguards keep ads out of "emotionally vulnerable, sensitive, or otherwise unsuitable conversations," per BleepingComputer.
The inventory problem is arithmetic
Digiday quotes Zeno Group's Shamsul Chowdhury on what the update does not touch. "There are about 2.5 billion prompts globally happening each day on ChatGPT, whereas there's about 15 billion searches on Google so the volume is inherently smaller," Chowdhury said.
- ChatGPT prompts2.5 billion a day
- Google searches15 billion a day
Source: Digiday, accessed 2026-10-05 (figures are Chowdhury's estimates)
The gap is wider than the ratio, Digiday reports, because ChatGPT carries one ad placement where a Google results page can carry several. An image-generation slot adds a second surface, but only for people who are generating images.
The company's own record on this
The ads sit beside a stretch of governance news at the same company. OpenAI's GPT-6.1 Sol system card drew attention for its cyber rating, and the departure of a safety-report lead, covered in the Robinson essay piece, turned on culture rather than product. Neither is about advertising, and neither outlet links them to the ad programme.
OpenAI has not published ad revenue, an advertiser count or a price for the image format. It has not said which countries come after the US.
The test is due to begin later in October, per both outlets, with no start date given.
Sources
More in A.I.
- 01Thinking Machines Ships Inkling 2 Mini, a 26.4B Apache 2.0 Model Sized for One DGX SparkThe model card lists 3.4B active parameters, a 1M-token window and a 20GB checkpoint, but does not say who ran most of its benchmark rows.
- 02Anthropic Closed Its Debut at $97.17, 17% Over the $83 IPO Price, as Nasdaq Fell 2%The $81.34 billion offering opened 34% higher, then gave back most of the gain while Nvidia, AMD and Micron dropped.
- 03Google's Gemini Tool Retrieval Preview Adds 26 Tokens Per Idle Tool and Publishes No Savings FigureAnthropic and OpenAI each reported token reductions for the same defer_loading flag. Google's November 19 post reports none.
- 04OpenAI Says New Safeguards Cut Reasoning Extraction From 54.0% to 0.2%The figures are from OpenAI's own red-team evaluation, and the post names DeepSeek and Moonshot AI as distillers of U.S. models.