top of page

ChatGPT Ads: Costs, Targeting and Ad Library Guide

Advertising inside ChatGPT is creating a new category of digital marketing: conversational advertising.


Unlike traditional search ads, which are triggered mainly by keywords, ChatGPT ads can be matched to the broader context and intent of a conversation. This means brands can reach potential customers when they are researching a problem, comparing products, or preparing to make a decision.


OpenAI began testing ads in the United States on February 9, 2026. Ads may appear for eligible users on ChatGPT’s Free and Go plans, while Plus, Pro, Business, Enterprise and Edu accounts remain ad-free. (OpenAI Help Center)


Alongside OpenAI’s official Ads Manager, an independent website called ChatGPT Ads Library now records sponsored placements observed inside ChatGPT. The database gives marketers an early way to research advertisers, creatives and the prompts associated with specific ads—but its results must be interpreted carefully.


What are ChatGPT ads?


ChatGPT ads are paid placements displayed below relevant ChatGPT responses.


A typical ad can include:

  • The advertiser’s name

  • A logo or favicon

  • A headline

  • Descriptive copy

  • A creative image

  • A landing-page link


OpenAI says ads are labelled as sponsored and visually separated from ChatGPT’s organic response. It also says advertisers cannot pay to influence, alter or rank the answer generated by ChatGPT. (OpenAI Help Center)


This distinction matters. An advertisement may appear because it is relevant to the conversation, but its presence does not mean that ChatGPT recommends or endorses the advertiser.


During the current test, a single ad unit may contain one or more products from one advertiser or products from multiple advertisers. Ads do not appear in Temporary Chats, near users identified as being under 18, or around certain sensitive conversations. (OpenAI Help Center)


How ChatGPT ad targeting works


ChatGPT ads are primarily selected using the context and apparent intent of the current conversation.


For example, someone asking how to improve payment security for an online store might see an advertisement for payment-processing software, fraud detection or cybersecurity services.


Advertisers can provide context hints describing the conversations, subjects or keywords where their offer may be useful. However, OpenAI says these hints are not conventional exact-match keywords and do not guarantee that an advertisement will appear for a specific prompt. (OpenAI Help Center)


Depending on the user’s settings, ad selection may consider:

  • The subject of the current chat

  • General location and language

  • Previous interactions with ads

  • Past chats

  • Saved ChatGPT memories

  • Other context used to personalise ChatGPT’s response


Users can turn off personalised advertising. When personalisation is disabled, ads can still be selected using the context of the current conversation, but past conversations, memory and ad history are not used. (OpenAI Help Center)


Can advertisers see ChatGPT conversations?


OpenAI says advertisers do not receive users’ chats, chat histories, memories, names, email addresses, precise locations, IP addresses or sensitive personal information.


Advertisers receive aggregated, non-identifying performance data such as total ad impressions and clicks. (OpenAI Help Center)


The practical distinction is that ChatGPT’s internal advertising system may analyse the context of a conversation to select an ad, but the underlying conversation is not disclosed directly to the advertiser.


How advertisers buy ChatGPT ads


Campaigns are created and managed through OpenAI’s Ads Manager.


Advertisers can set campaign objectives, budgets and bids, create ad groups, submit creatives, add context hints and review performance results. (OpenAI)


ChatGPT Ads currently supports two principal buying models:


Cost per click


With a Clicks campaign, the advertiser pays for valid clicks.


OpenAI recommends beginning with a maximum bid of approximately US$3 to US$5 per click. This is bidding guidance—not a guaranteed price or a published market-average CPC. The amount an advertiser actually pays may differ based on competition, relevance and campaign conditions. (OpenAI Help Center)


Cost per thousand impressions


With a Reach campaign, the advertiser pays for every 1,000 valid impressions.


OpenAI uses a relevance-weighted, second-price auction. Advertisements are evaluated using both the advertiser’s bid and their relevance to an eligible conversation. A higher bid may therefore lose to another advertisement if the competing ad is substantially more relevant. (OpenAI Help Center)


Ads Manager Beta currently reports metrics including:

  • Impressions

  • Clicks

  • Spend

  • Click-through rate

  • Average CPC

  • Average CPM

  • Conversions


Advertisers can also add UTM parameters to landing-page URLs and configure conversion measurement to evaluate actions taken after an ad click. (OpenAI Help Center)


Where is ChatGPT Ads Manager available?


As of July 16, 2026, OpenAI lists Ads Manager as available in:

  • Australia

  • Canada

  • Japan

  • South Korea

  • New Zealand

  • United Kingdom

  • United States


India is not currently included in the availability list. Businesses outside supported markets can register their interest, but OpenAI has not published a confirmed date for self-serve access in India. (OpenAI Help Center)


This country list refers specifically to advertiser access to Ads Manager. OpenAI may expand advertiser and consumer availability gradually as the testing programme develops.


What types of ChatGPT ads are allowed?


OpenAI’s initial advertising programme focuses mainly on categories such as:

  • Lifestyle and household products

  • Local services

  • Travel and experiences

  • Digital products

  • Education


Some financial, healthcare and medical advertisers may be approved on a case-by-case basis, particularly in the United States. These advertisers may need to provide licences or other evidence of eligibility. (OpenAI)


OpenAI’s policies currently prohibit or restrict several categories, including:

  • Political advertising

  • Gambling

  • Alcohol and tobacco

  • Dating and adult services

  • Weapons

  • Recreational drugs

  • Scams and fraudulent offers

  • Many legal services

  • Unsupported medical or wellness claims


Ads are also not supposed to appear near sensitive contexts involving mental or personal health, emotionally vulnerable interactions, self-harm, politics, fraud, violence and similar high-risk subjects. (OpenAI)


What is the ChatGPT Ads Library?


ChatGPTAdLibrary.com is a public ad-intelligence database that records advertisements observed inside ChatGPT.


It is important not to confuse this database with OpenAI’s own products:

  • OpenAI Ads Manager is used by advertisers to run and measure campaigns.

  • ChatGPT Ads History lets eligible users review advertisements they personally viewed or interacted with.

  • ChatGPTAdLibrary.com collects and organises advertisements observed through its own research process.


The library says it repeatedly probes ChatGPT using realistic commercial prompts across different industries. When ChatGPT returns a sponsored placement, the library records:


  1. The advertiser

  2. The ad creative and copy

  3. The prompt associated with the placement


It also says each observation is stored with a timestamp and the underlying response used as evidence. (chatgptadlibrary.com)


At the time of review on July 16, 2026, the site reported:

  • 9,046 unique ad creatives

  • 2,737 observed advertisers

  • 92,091 placements

  • 109 niches


These are the library’s own sampled figures—not official totals published by OpenAI. (chatgptadlibrary.com)


Is the ChatGPT Ads Library complete?


No.


The website’s homepage promotes the service as a way to see “every ad” and “every advertiser.” However, its methodology page explicitly states that the dataset represents only a subset of everything running inside ChatGPT. Coverage is strongest in industries and niches that the site probes most frequently. (chatgptadlibrary.com)


The library does not have access to OpenAI’s internal campaign, impression, targeting or expenditure data.


A placement in the library means the advertisement was observed in response to one of the prompts tested by the site. It does not reveal:

  • The advertiser’s total campaign budget

  • Actual platform-wide impressions

  • Actual expenditure

  • Complete geographic targeting

  • All creatives used by the advertiser

  • The advertiser’s private context hints

  • The total number of competing advertisers


The library says advertiser presence and prompt coverage can serve as proxies for advertising activity, but they should not be treated as direct measures of budget or total market share. (chatgptadlibrary.com)


How marketers can use the ChatGPT Ads Library


Despite its limitations, the library can provide useful directional intelligence.


1. Identify early advertisers


Marketers can examine which brands are already appearing in their industry and which categories appear relatively crowded or underdeveloped.


This can help businesses assess whether competitors are experimenting with conversational advertising before the channel becomes more mature.


2. Study ad positioning


The database can reveal the types of headlines, benefits and calls to action advertisers use.


Marketers can analyse whether competitors focus on:

  • Lower pricing

  • Ease of use

  • Trust and security

  • Speed

  • Product comparisons

  • Free trials

  • Consultations

  • Discounts


This does not reveal which creative performs best, but it provides evidence of how brands are positioning themselves.


3. Examine prompt-to-ad relationships


One of the library’s most distinctive features is its association between an observed ad and the prompt used when the placement appeared.


This may help marketers understand the types of conversational intent connected to an advertiser. However, an observed prompt is not necessarily the advertiser’s actual targeting instruction. ChatGPT considers several matching signals, and the library does not have access to the advertiser’s private Ads Manager settings.


4. Find content and positioning gaps


A business can compare the questions people may ask during the buying journey with the advertisements already appearing around those questions.


For example, a cybersecurity company might investigate prompts related to:

  • Protecting customer data

  • Passing a security audit

  • Preventing phishing

  • Selecting penetration-testing tools

  • Meeting compliance requirements


The company could then evaluate whether existing advertisers answer those needs effectively or leave important gaps.


5. Build a creative swipe file


Marketing teams can maintain a structured record of:

  • Advertiser

  • Category

  • Prompt

  • Headline

  • Offer

  • Call to action

  • Landing-page type

  • Creative format

  • Date observed


This turns individual ad examples into a repeatable research system rather than an unorganised collection of screenshots.


What ChatGPT ads change for digital marketing


ChatGPT advertising changes the unit of targeting.


In search advertising, marketers typically target a query or keyword. In social advertising, they often target an audience profile or behaviour.


ChatGPT ads introduce a third approach: targeting the context of an active decision-making conversation.


A user may explain:

  • What they are trying to achieve

  • What problems they are facing

  • Which alternatives they are considering

  • Their budget or constraints

  • How urgently they need a solution


That context may help the advertising system identify a relevant commercial message more precisely than a short search query.


The risk is that conversational relevance can feel more influential than conventional advertising. Clear labelling, separation from organic answers, meaningful user controls and independent transparency therefore become particularly important.


A practical strategy for advertisers


Businesses considering ChatGPT ads should begin with a narrow test rather than copying an existing Google Ads or Meta Ads campaign.


A strong initial process would be:

  1. Identify a specific customer problem.

  2. List the questions customers ask before purchasing.

  3. Group those questions by intent.

  4. Create context hints around each intent group.

  5. Write ads that directly address the user’s situation.

  6. Send users to a closely matched landing page.

  7. Add UTM parameters and conversion tracking.

  8. Measure conversion quality, not only clicks.

  9. Review observed competitor activity through the ChatGPT Ads Library.

  10. Update targeting and creative based on actual results.


Because the format is conversational, vague brand messaging is likely to be less useful than a clear explanation of why the product is relevant at that particular moment.


Frequently asked questions


Are ChatGPT ads part of the answer?

No. OpenAI says advertisements appear below the response, are labelled as sponsored and remain separate from the answer. (OpenAI Help Center)

OpenAI says advertisers cannot shape, rank or alter ChatGPT’s responses. An advertisement appearing near an answer should not be interpreted as a recommendation or endorsement. (OpenAI Help Center)

OpenAI says advertisers do not receive users’ conversations, memories or personal information. They receive aggregated advertising-performance data. (OpenAI Help Center)

It should not be confused with OpenAI’s Ads Manager or ChatGPT’s account-level Ads History. Its database is created by probing ChatGPT and recording observed sponsored placements rather than accessing OpenAI’s complete internal campaign data. (OpenAI Help Center)

No. Its methodology states that its dataset is a sample and that coverage varies according to the industries and prompts it tests. (chatgptadlibrary.com)

India is not currently listed among the markets where advertisers can access OpenAI Ads Manager. No confirmed Indian self-serve launch date has been published. (OpenAI Help Center)


Conclusion


ChatGPT ads represent a significant shift from keyword-based advertising towards context-based, conversational advertising.


OpenAI’s advertising system uses conversation intent, creative relevance and advertiser bids to choose eligible ads. Advertisers can buy campaigns using CPC or CPM objectives, while users retain controls over personalisation and ad history.


The independent ChatGPT Ads Library adds a useful research layer by recording observed advertisers, creatives and associated prompts. But its database is a sampled view of the market—not a complete or official record of all ChatGPT advertising activity.


For marketers, the opportunity is not simply to place another display ad. It is to understand what a potential customer is trying to accomplish and present a relevant, credible offer at the right point in that decision-making process.


The article is structured for a company blog or SEO publication and includes the ChatGPT Ads Library as a substantial, clearly qualified section.


 
 
 

Comments

bottom of page