ChatGPT Visual Ads Begin This Month, Not for Plus or Pro

OpenAI starts testing visual ads in ChatGPT in the US later this month, placing sponsored product images beside AI-generated pictures. It is the company's second ad format in eight months, and the first one that competes for attention with the output itself.

The Verge reported the launch on Monday, and OpenAI described the same move as a new visual ad format alongside expanded measurement tools, attribution partnerships and brand suitability work for advertisers.

What actually changes in the product

Until now, ChatGPT ads were text. According to The Verge, they showed only "a company's name, logo, and a link to its product" in a sponsored section beneath the chat. The new format adds imagery, and it will "initially" appear when you generate images.

Two constraints are worth reading closely. The ads stay separate from the image you generated, and OpenAI says they will "not influence the answers ChatGPT provides." TechCrunch reports the same limits and adds that the ads will be clearly labeled and come from an initial test group of advertisers.

OpenAI also says it has guardrails that decide when a conversation is "appropriate" for ads, with the stated goal of keeping them out of "emotionally vulnerable, sensitive, or otherwise unsuitable contexts." Neither the announcement nor the coverage of it describes how that classifier decides, which is the part worth tracking once the test is live.

Ad surface	Status	Where
Text ads under the chat	Live since February	US, plus India from August
Visual ads beside image results	Testing later this month	US only, initial advertisers

Is a paid plan now the ad-removal plan?

For anyone weighing a subscription, this is the practical question. The Verge reports plainly that ads will not appear for subscribers on ChatGPT's Plus, Pro or Enterprise plans. That turns the ad rollout into a feature comparison rather than a policy debate: the free tier buys you ads, the paid tiers do not.

TechCrunch makes the same point from the business side, writing that ads "could prove disruptive to the user experience" and that pushing casual users toward a subscription may be partly the point. That framing is the outlet's analysis rather than an OpenAI statement, but it matches the structure of the product — ads land exactly where no one is paying.

Worth separating two decisions here. If you already pay for ChatGPT, nothing changes this month. If you are on the free tier and generate images regularly, the honest comparison is not ads versus no ads; it is whether image generation with sponsored results alongside is still the fastest path for your workflow, against other image tools you could use for free.

The measurement stack is the real announcement

The visual format grabbed the headlines, but the longer list in OpenAI's announcement is infrastructure for advertisers. Per TechCrunch, the company is expanding click attribution through AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge and Tenjin, and adding full-funnel measurement partners Fospha, Measured and INCRMNTAL.

On top of that, OpenAI has partnered with Haus, Measured and WorkMagic for "geo-based" ad experiments, and is developing brand suitability evaluation pilots with DoubleVerify and Integral Ad Science.

That is a conventional ad-tech stack, assembled quickly. Attribution partners let advertisers prove a click turned into a purchase; brand suitability vendors let them verify where their logo appeared. Neither is glamorous, and both are what large advertisers require before moving real budget. The shape of the announcement suggests OpenAI is courting the buyers who ask those questions, not just testing a layout.

Why the timing looks defensive

TechCrunch places the move after the launch of Meta's Muse, an agentic AI app the outlet describes as free to everyone and a growing competitor to leading chatbots. OpenAI introduced ads earlier this year to support its free and low-cost tiers, and expanded them to India in August.

Scale is the argument for advertisers. TechCrunch notes that once the ads roll out globally, they will eventually reach ChatGPT's 1.2 billion weekly users — a number that applies to the eventual rollout, not this month's US test with an initial advertiser group.

Read together, the sequence is familiar from consumer internet history: a free tier gets expensive to run, a competitor makes free table stakes, and the ad format grows from a text link into something you can see. February's version sat below the conversation. This one sits beside the thing you came for.

What to watch next

• Where the second surface lands. OpenAI says it plans to try out new formats, and image generation is a deliberately low-stakes starting point. Search-style and shopping queries are the obvious next candidates.
• Whether the guardrail holds. The "appropriate for ads" classifier is the single claim in this announcement that users can check in practice, and the first screenshot of an ad in a sensitive thread will test it.
• Whether anyone reports conversions. With this many attribution partners signed up, the first advertiser case studies will say more about ChatGPT as an ad channel than any announcement does.

For advertisers, the practical step this month is narrow: the test group is closed, so the work is deciding whether your measurement vendor is on the list above. For everyone else, the thing to notice is that the free tier of a 1.2 billion-user assistant now has a visible price, and it is paid in attention.