OpenAI has officially begun testing ads inside ChatGPT in the United States, marking a major shift from AI as a subscription product to AI as an ad‑supported consumer platform for at least a portion of its user base. The test is limited to logged in adult users on the Free and Go tiers, while Plus, Pro, Business, Enterprise, and Education remain ad‑free.
At the same time, early industry reporting (and OpenAI confirmation to Adweek) indicates the initial advertiser access is enterprise gated with a minimum commitment around $200,000, and pricing reportedly around $60 CPM. This positions the program as a premium, tightly controlled pilot, not a self‑serve channel marketers can test the way they can on Meta or Google.
Marketers: Here’s what you need to know about OpenAI’s new ads program.
What it is
OpenAI is testing ads in ChatGPT as a way to support broader access to AI, particularly for the Free tier and the low‑cost Go plan, without changing the core behavior of ChatGPT answers. OpenAI frames ads as funding infrastructure and product investment while preserving user trust through strict separation and user controls.
These are not sponsored answers. OpenAI explicitly states:
- ads do not influence answers,
- advertisers cannot shape/rank/alter responses, and
- ads are separate and clearly labeled.
This distinction matters because the credibility of conversational interfaces depends on users believing the assistant is optimizing for helpfulness, not revenue.
Placement and experience
During the pilot:
- Ads can appear below the end of a response and are clearly labeled and separated from the answer.
- Ads currently do not appear in:
- Temporary Chats
- when logged out
- after generating an image
- in the ChatGPT Atlas browser (for now)
- Temporary Chats
OpenAI also notes that you may typically see a single ad unit below a response when there is a relevant match, with overall context considered in longer threads.
Targeting: contextual first, optional personalization
OpenAI describes ad selection as primarily driven by what you’re discussing in your current chat thread. If the user opts into personalized ads, OpenAI may also use past chats, memory, and prior ad interactions (hiding or engaging with ads) to improve relevance.
Crucially, OpenAI says the signals used to select ads stay within ChatGPT and are not shared with advertisers, and advertisers never receive things like chats, name/email, precise location, IP address, or sensitive information.
Ad Controls
Free and Go users in the U.S. can:
- turn off ad personalization,
- dismiss/hide ads and give feedback,
- see why an ad was shown,
- and clear ads-related data.
If a user clears ad data, OpenAI states it may be retained for up to 30 days before being removed from servers.
Going Ads Free
If users prefer not to see ads, OpenAI offers an Ads Free option on the Free plan, but with lower usage limits and reduced feature access. This is presented as a direct trade: ads support more access; opting out reduces your daily free usage.
The advertiser reality right now
Adweek reports that OpenAI has confirmed select advertisers are being asked to commit at least $200,000 as it rolls out the beta, described as small and tightly controlled, while OpenAI evaluates which ad formats provide value to users.
EMarketer further summarizes proposals reportedly ranging from $100K–$250K (per Adweek), reinforcing that the current state is enterprise and negotiated.
CPM and buying model
Multiple industry sources report pricing around $60 CPM and that the pilot is sold on impressions rather than clicks.
For most marketers, that means ChatGPT ads are not yet a viable performance channel. They are currently closer to a Super Bowl advertisement than they are to Google Ads.
Measurement constraints
OpenAI’s help documentation states that advertisers receive aggregated reporting, such as views and clicks (and that OpenAI may explore additional measurement insights later).
The Verge’s reporting, citing The Information, similarly notes that early advertisers may not receive the same depth of action-level measurement as with Google/Meta due to OpenAI’s privacy commitments.
Why does this matter?
Even if your brand will not participate in a $200K+ pilot, ChatGPT ads change the marketing landscape in three immediate ways:
Conversation intent becomes a targetable moment
Ads are matched to what people are actively discussing, often at the exact moment they are comparing options, planning, or making decisions.
This is different from keyword search intent: the user’s constraints and context are often richer and more explicit.
Trust is the constraint
OpenAI is explicitly designing the system around answer independence and “don’t break trust.”
That means ad creative and landing experiences that feel spammy, bait‑and‑switch, or irrelevant will likely underperform and/or be throttled by user controls and feedback loops.
Organic AI visibility remains a prerequisite
Because ad inventory is limited and the pilot is gated, most brands will compete in the near term through:
- AI discoverability (how often your brand appears in assistant-driven research and comparison),
- authority content (clear claims, FAQs, proof points),
- and machine-readable product/service data (structured information that assistants can reliably interpret).
- This is exactly why AI SEO / GEO / LLM visibility is becoming a core discipline.
What CVAC is doing to prepare (and what clients should do now)
CVAC preparation plan
- Register and pursue advertiser access
OpenAI’s “Advertise with ChatGPT” page is currently the clearest entry point to express interest and stay informed as buying models expand. - Maintain a live research brief on ChatGPT ads
The ecosystem is moving quickly (controls, formats, pilot economics). We are tracking confirmed facts vs. reported details so clients can plan with precision. - Execute AI visibility + conversion readiness for clients
Ads are only one lever. Whether or not you buy them, you need to be:
- easy to cite and verify in assistant responses,
- credible in decision moments,
- and optimized for “conversation-led” journeys (problem → options → decision).
- easy to cite and verify in assistant responses,
- Schedule an AI SEO/GEO Visibility Audit to understand how visible your brand is across ChatGPT and other assistant-led discovery flows.
- Schedule an AI Marketing Strategy Session: align content, technical SEO, measurement, and channel mix for 2026.
- Download the PDF version of this briefing.
- Read our Atlas white paper and related articles on assistant-first discovery and commerce.
TLDR
- Ads are live in test (U.S.) for Free + Go users who are logged in adults; no ads on Plus/Pro/Business/Enterprise/Edu.
- Ads appear below the end of a ChatGPT response, clearly labeled sponsored, and visually separated from the answer.
- OpenAI says ads do not influence ChatGPT answers and run on separate systems from the chat model.
- Advertisers do not see chats, memories, or personal details; reporting is currently aggregated views/clicks (with the possibility of more measurement later).
- Minimum advertiser commitment: OpenAI confirmed to Adweek that select advertisers are being asked for at least $200,000 in the early test; some pitches reportedly go higher.
- Pricing: widely reported at $60 CPM, sold on impressions.
- Sensitive-topic controls: ads are not eligible near sensitive/reg topics (health, mental health, politics), and certain regulated advertiser categories are excluded at launch.