ChatGPT will start showing picture ads while you generate images

The new format uses images instead of only text. OpenAI says the visuals can show product inspiration, how a product is used, or the experiences a service makes possible, as BleepingComputer and Unite.AI report.

"Initially, we'll test this new ad format during image generation in ChatGPT," OpenAI said, as quoted by BleepingComputer. "Ads will be clearly labeled, and remain separate from the image being created."

The details so far:

• When: testing begins later in October 2026.
• Where: the US only, with an initial group of advertisers.
• Placement: during image generation, not inside the image itself.
• Labeling: OpenAI says the ads will be clearly marked as ads.

OpenAI also repeated a promise it has made since the start: ads don't influence ChatGPT's answers.

For the past seven months, advertisers have worked with a fairly simple format, Digiday reports. It includes the advertiser's name, logo, headline, description, landing page and an image.

Those ads appear alongside ChatGPT's answers. At launch in February, MacRumors reported that OpenAI shows ads based on the conversation topic, past chats and past interactions with ads. A chat about recipes, for example, might surface an ad for meal kits.

The new visual format gives advertisers a different pitch. Instead of describing a product in a line of text, they can show what it looks like in use.

Not everyone. Ads are limited to the Free and Go tiers, and only for logged-in adults, according to MacRumors' report on the February launch.

Ad-free plans: Plus, Pro, Business, Enterprise and Education subscribers don't see ads.

No ads for minors: OpenAI said it doesn't plan to show ads to users under 18.

Sensitive topics: ads aren't shown in conversations about health, mental health or politics, MacRumors reported.

ChatGPT now reaches 1.2 billion people every week, OpenAI says. BleepingComputer notes that ads give OpenAI a way to make money from the many users who don't pay for a subscription.

The other half of Monday's announcement is about proof. Advertisers want to know whether a ChatGPT ad led to a sale, and until now that's been hard to show.

OpenAI is expanding its pixel and Conversions API, which measure actions such as purchases, leads, app installs and sign-ups, Unite.AI reports. It's also adding a long list of outside partners:

• Conversion data: integrations with Hightouch, Tealium and LiveRamp let advertisers send results from their own systems.
• Attribution: partners include AppsFlyer, Adjust, Branch, Triple Whale, Kochava, Northbeam, Singular, Airbridge, Tenjin and DV Rockerbox.
• Wider measurement: Fospha, Measured and INCRMNTAL handle full-funnel measurement.
• Early experiments: OpenAI is testing geo-based incrementality with Haus, Measured and WorkMagic, and brand impact studies with Kantar and Cint.

Advertisers can also count conversions that happen within a day of someone seeing an ad, even without a click, according to Unite.AI.

OpenAI shared some early numbers from its partners, Unite.AI reports.

• WeightWatchers: its cost per acquisition on ChatGPT ads was 15.3% lower than its blended paid search benchmark, according to DV Rockerbox.
• Dose: WorkMagic reported that 67% of incremental purchases for the wellness brand came from new customers.
• Portland Leather: 93% of its visitors from ChatGPT ads were new, according to Triple Whale.

OpenAI also said that in an early analysis, 52.7% of eligible one-day view-through conversions happened within an hour of the ad being shown.

These are figures chosen by OpenAI and its partners, not independent audits.

OpenAI says its placement rules try to keep ads out of conversations that are emotionally vulnerable, sensitive or otherwise unsuitable. It uses automated review, human oversight and ongoing monitoring, according to Unite.AI.

New outside checks: DoubleVerify and Integral Ad Science will run pilots to evaluate how OpenAI applies its brand safety standards. Both will work in a controlled testing environment, without access to private user conversations, Digiday reports.

"Advertising in ChatGPT differs from web and video placements," DoubleVerify CEO Mark Zagorski said, in Unite.AI's summary, because the conversation itself, not a page or a video, decides where an ad appears.

For brands: a feature called Negative Phrases lets qualifying advertisers keep their ads away from specific kinds of content.

Advertisers have had three main complaints since February, Digiday reports: they can't measure the ads properly, there aren't enough formats, and there isn't enough inventory for what they want to spend.

Monday's update tackles the first two. The third is harder. "There are about 2.5 billion prompts globally happening each day on ChatGPT, whereas there's about 15 billion searches on Google," Shamsul Chowdhury, an executive at Zeno Group, told Digiday. He added that ChatGPT has only one ad placement, while Google has several.

The timing is deliberate too. The announcement lands at the start of Advertising Week New York, where OpenAI executives plan to get feedback from marketers, according to Digiday.

The ad business has grown quickly this year.

• February 9: ads began for US users on the Free and Go tiers, MacRumors reported.
• August: ads expanded to Europe, and OpenAI said it would start showing them in India with 50 brands, TechCrunch reports.
• October: the visual format test begins in the US.

In India, OpenAI partnered with the agencies WPP and Omnicom and said it would launch an ad manager for marketers, with a minimum daily budget of ₹725 ($7.60), TechCrunch reports. "With ChatGPT Ads, businesses of every size can introduce themselves at relevant, high-context moments when decisions are beginning to take shape," said Dave Dugan, OpenAI's head of global ads solutions.

The push comes as OpenAI builds up its revenue. TechCrunch, citing The Wall Street Journal, reports OpenAI recorded $6.7 billion in revenue in the quarter ended June 2026, up from $5.7 billion the quarter before.

Not every rival is following. When ChatGPT ads launched, Anthropic promised to keep Claude ad-free, MacRumors reported, and ran a Super Bowl commercial saying a conversation with Claude isn't the right place for ads.

OpenAI has been adding shopping features to ChatGPT too, including a recent virtual try-on tool for clothes.

• Free and Go users: you may start seeing picture ads while ChatGPT generates images, if you're in the US and in the test.
• Your images shouldn't change: OpenAI says the ads stay separate from what you're creating.
• Paid plans stay clean: Plus, Pro, Business, Enterprise and Education plans don't show ads.
• You have options: you can turn off ad personalization, delete ad data, or trade fewer free messages for no ads, according to MacRumors.