ChatGPT ads come to image generation in a US test

OpenAI tests a visual ad format in ChatGPT image generation

OpenAI is adding an ad format built around pictures, and it will test it first in ChatGPT image generation. The company says ChatGPT reaches 1.2 billion people each week and that ads let it serve far more people than subscriptions alone.

The advertising test starts in the US in late October

Testing starts later in October 2026 in the US with an initial group of advertisers. OpenAI announced the format on October 5, so "later this month" means the second half of October.

The announcement names no other country for this test. Advertisers can sign up at ads.openai.com.

Product ads appear while the image generator works

The first place you will see the new ads is the screen shown while ChatGPT creates an image. You ask for a picture, and while it renders, an ad can sit on the same screen.

The announcement does not name regular text chats as a surface for this format. Ads in regular chats still appear below the end of a response.

Visual ads stay apart from your ChatGPT photo

OpenAI says the ads are clearly labeled and remain separate from the image being created, so a brand will not show up inside your picture. That answers the obvious worry about ads during image generation.

Labels on every ad and privacy for your chat history

Every ad is labeled, and advertisers never see your conversations. OpenAI's Help Center says ads run on systems separate from the chat model, so advertisers cannot change, rank or shape ChatGPT's answers.

Advertisers have no access to your chats, chat history, memories or personal details. They get only aggregated numbers about their own ads, such as total views and clicks.

OpenAI's grocery example shows shopping-style product ads

The pictures in these ads show product ideas, how a product is used, or the experience it makes possible. OpenAI says this helps people discover a brand, see what sets it apart and decide what to buy.

The sample image on OpenAI's announcement page shows a grocery brand called Heirloom Grocery. Below the image-generation progress indicator sits a row of vegetable and meal photos, an "Ad" label and a "Learn more" button.

Who sees ads: ChatGPT Free and Go, not Plus or Pro

The visual ad test covers users on the Free and Go plans, and Plus and Pro accounts do not get ads. Your plan decides whether you see them.

The ChatGPT Go plan is the only paid plan with ads

Go, OpenAI's low-cost tier between Free and Plus, is the one paid plan that shows ads. GPT-5 in ChatGPT: Pricing, Free Tier, Models compares what each tier includes.

OpenAI's Help Center lists Plus, Pro, Business, Enterprise and Edu as ad-free. One detail is easy to miss: the official Japanese edition of the announcement leaves out the plan wording, so only the English original states that this test runs on Free and Go.

No ads under age verification or in temporary chat

Even on Free and Go, ChatGPT does not show ads in three situations. Being on an eligible plan is not enough on its own.

• Accounts that OpenAI identifies as belonging to people under 18
• Temporary Chat
• Conversations that touch on personal health, mental health or politics

The October announcement adds that placement guardrails check each conversation and keep ads out of emotionally vulnerable or sensitive contexts. Automated review, human oversight and ongoing monitoring catch policy violations.

ChatGPT ads in Japan and the UK keep the regular format

Outside the US, ChatGPT shows the regular ad format, and the visual format has no announced date. The October 5 post only describes the US test.

ChatGPT Ads reached Japan and the UK on August 11

On August 11, 2026, OpenAI updated its ads test post to say ChatGPT Ads had launched in the UK, Mexico, Brazil, Japan and South Korea. The test began in the US on February 9, 2026, and had expanded to Canada, Australia and New Zealand before the August update.

Several reports describe those regular ads as text-only, but they are not. OpenAI's advertiser guide says every ad includes a brand name, logo, title, copy, landing page and a square image.

The new format differs in two ways. It appears during image generation, and it leads with product photos.

Two ChatGPT settings remove ads, and one has a message limit

You can remove ads by moving to Plus or Pro, or by switching the Free plan to Ads-Free, which costs nothing but takes features away. The right choice depends on whether you still want to make images.

Upgrade to ChatGPT Plus or Pro

Plus and Pro show no ads, and this is the only ad-free route that keeps image generation. It is also the route for Go users, because Go cannot switch to Ads-Free.

Ads-Free removes deep research and lowers the message limit

Ads-Free removes ads on the Free plan, but you get fewer messages a day and lose tools such as image generation and deep research. If you want to make images, this option does not fit. Ads-Free users would not see the visual ads anyway, because those appear only during image generation.

To switch, follow the steps in OpenAI's Help Center. The menu names below are the ones it uses.

1. Open Settings in ChatGPT
2. Select Ads controls
3. Tap Change plan to go ad-free
4. Choose Reduce message limits
5. Confirm the change

Turning off personalization in settings keeps ads on

Turning off ad personalization does not remove ads. It changes how ads are picked.

With personalization off, ChatGPT stops using your other chats, ad history and topics to choose ads. You still see ads based on the chat you have open.

Brand safety and measurement for ChatGPT advertising

The other half of the announcement is for advertisers: more ways to measure results and more controls over where ads appear. It does not change your screen, but it shapes which ads you get and where they do not appear.

Ten attribution partners and conversion API support

OpenAI now works with ten attribution partners that tie ad clicks to sales using the Conversions API, reporting and click attribution. It also named partners for sending conversion data and for testing incrementality.

Incrementality means the extra sales an ad causes compared with no ad. OpenAI also shared early numbers that its partners reported:

• DV Rockerbox: WeightWatchers paid 15.3% less per customer acquired through ChatGPT Ads than its average across paid search channels
• WorkMagic: 67% of Dose's incremental purchases came from net-new customers
• Triple Whale: 93% of Portland Leather's visitors from ChatGPT Ads were new

These are results for single advertisers, not averages across ChatGPT Ads. Each partner measured its own client with its own method.

DoubleVerify and Integral Ad Science test brand suitability

OpenAI is building brand suitability pilots with DoubleVerify and Integral Ad Science that never give them access to private conversations. The partners assess how OpenAI applies its brand safety standards in a controlled test environment.

Qualifying advertisers can also use Negative Phrases to keep their ads out of contexts that clash with their own policies. This is a control on the advertiser side, separate from OpenAI's own guardrails.

Shopping ads start in the US, and chat privacy stays intact

Visual ads start in US image generation for Free and Go users in late October, and they stay apart from the images you make. Other markets keep regular ads for now.

If you generate images often and do not want ads, Plus or Pro is the practical choice. Ads-Free takes away image generation, so it does not suit that use.