Image: OpenAI

OpenAI is rolling out a new visual ad format in ChatGPT and expanding how advertisers measure results. The company says the ads help people picture products and services in real life, while measurement partners and brand-safety pilots aim to make the channel feel more like a mature media buy.

ChatGPT now reaches 1.2 billion people each week, OpenAI says, and ads help fund free and low-cost access beyond subscriptions alone. The new format is part of that push — but the bigger story for marketers is whether ChatGPT Ads can finally prove outcomes the way Google and Meta do.

OpenAI will first test the visual ads during image generation in ChatGPT. Ads will be clearly labeled and kept separate from the image being created. As with other ChatGPT ad formats, OpenAI says advertising does not influence the answers ChatGPT provides. Testing starts later this month in the US with an initial group of advertisers.

On the measurement side, OpenAI is adding conversion-data integrations with Hightouch, Tealium, and LiveRamp, plus attribution and reporting partners including AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge, and Tenjin. Full-funnel partners include Fospha, Measured, and INCRMNTAL. Early incrementality work with Haus, Measured, and WorkMagic explores geo-based experiments. OpenAI is also piloting brand-suitability evaluation with DoubleVerify and Integral Ad Science without giving those partners private user conversations, and Negative Phrases are now available to qualifying advertisers.

That sits on top of a fast ads rollout. OpenAI began testing ChatGPT Ads in the US for logged-in adults on Free and Go plans. Plus, Pro, Business, Enterprise, and Education plans stay ad-free. From the start, OpenAI pledged that answers stay independent, conversations stay private from advertisers, and it does not sell user data to advertisers. Ads later expanded across Europe, Asia, and other regions, and OpenAI has said the product is now available in more than 60 countries after reaching a $1 billion annualized revenue run rate in under 200 days.

The arc matters: the platform went from a sponsored unit under answers to formats, pixels, conversions APIs, and now visual placements plus partner measurement. The privacy pitch is unchanged in this update — ads remain separate, and brand-safety pilots are framed as not reading private chats. OpenAI's help docs still say advertisers do not get chats, chat history, memories, or personal details — only aggregate performance stats such as views and clicks.

Google Search ads show near results when keywords match what someone typed on the Google Search Network. Meta's Advantage targeting can expand some audience options when that may improve results, while hard controls such as location, age, gender, and exclusions still apply. ChatGPT Ads sit in a different moment: a live conversation about a goal, not a keyword auction or a social feed. OpenAI still limits ads to Free and Go users, keeps answers off the ad system, and withholds chats from advertisers — a narrower inventory slice than Google or Meta, with a stronger "intent in progress" claim if measurement holds up.

The visual format during image generation is a natural next step. People are already looking at pictures ChatGPT made; a labeled product image nearby is closer to how discovery works in creative tools than a text card under a reply. Measurement catch-up is the real unlock. Until advertisers can trust attribution and lift the way they do on Google and Meta, ChatGPT Ads stay a test budget. This update is OpenAI trying to close that gap without rewriting its ads principles.