OpenAI has moved ChatGPT Ads from an early US pilot toward a broader self-serve advertising platform. On May 5, 2026, the company opened Ads Manager to all US advertisers and added cost-per-click bidding alongside its existing CPM buying model. The change gives advertisers a more familiar way to manage spend, but ChatGPT Ads remain fundamentally different from search advertising: relevance is guided by conversation context and advertiser-provided context hints, not exact-match keywords or demographic targeting.
The rollout matters because it begins to define how commercial messages may operate inside AI chat interfaces. Ads appear at the bottom of eligible ChatGPT conversations, are clearly labeled, and are designed to remain separate from the assistant's answers. OpenAI says advertising cannot influence what ChatGPT tells a user, a governance boundary that will be central to advertiser and user trust as the inventory, targeting, and measurement stack develop.
From pilot program to self-serve buying
OpenAI began distributing ChatGPT Ads in the US through a pilot in February 2026. HelloFresh was among the early participants, according to its February 9 press release. The initial format established the core experience: a distinct sponsored placement below a conversation rather than an ad embedded in or shaping the answer itself.
The subsequent self-serve expansion is the more consequential operational step. In OpenAI's announcement on new ways to buy ChatGPT Ads, the company described CPC bidding as an addition to CPM campaigns. That gives advertisers a choice between buying exposure and paying for engagement, while OpenAI continues to build the controls needed for a larger ad marketplace.
Early third-party reporting placed observed CPCs around $2 to $5. OpenAI's Help Center recommends starting CPC bids of $3 to $5 per click. Those figures should be treated as early guidance and reported market observations, not a promise of future campaign pricing. Actual outcomes can vary as inventory grows, more advertisers enter the auction, and campaign settings evolve.
| Buying or targeting element | ChatGPT Ads approach | What it means for advertisers |
|---|---|---|
| Pricing model | CPM and CPC bidding | Campaigns can be oriented toward impressions or clicks. |
| Keyword targeting | No exact-match keywords | Advertisers cannot recreate a conventional search-keyword strategy. |
| Relevance controls | Context hints at the ad-group level | Creative and context guidance become important inputs to matching. |
| Audience controls | No demographic targeting; geo-targeting by country or region | Planning must account for more limited targeting options. |
| Measurement | Platform metrics, pixel, and Conversions API | Advertisers can track delivery and connect campaigns to conversion measurement. |
Context, rather than keywords, is the operating model
The key strategic adjustment is that ChatGPT Ads are not a keyword-search product. OpenAI uses context-based matching, and advertisers provide context hints at the ad-group level to help guide relevance. The company does not offer exact-match keyword targeting or demographic targeting, although campaigns can be geo-targeted by country or region.
That approach requires a different planning discipline. Advertisers need to articulate the situations, needs, and adjacent topics in which their offer is useful, then ensure their creative makes that relevance understandable. A keyword list alone is not the campaign blueprint. Brands also need to test whether the context in which an ad appears produces meaningful downstream intent, rather than assuming that a click carries the same meaning it might in a conventional search campaign.
OpenAI has added capabilities including custom audiences and conversion-optimized campaigns as the product has matured. These additions point to a platform that is gaining more performance-oriented controls, but they also reinforce the need to treat current practice as iterative. The available research characterizes ChatGPT Ads as an actively evolving ecosystem, not a finished replacement for established search or social ad platforms.
Measurement and trust are the early constraints
Ads Manager provides delivery metrics including impressions, clicks, spend, click-through rate, average CPC, and average CPM. Advertisers can measure conversions through a pixel or the Conversions API. These are necessary foundations for evaluating performance, particularly for teams accustomed to optimizing against business outcomes rather than media delivery alone.
Still, measurement availability is not the same as mature benchmarking. Reported CPMs declined from launch levels as inventory expanded, illustrating why initial results may not remain representative. Businesses should establish their own baselines, keep test budgets controlled, and separate what is measured on-platform from what can be attributed to conversion activity.
The product's safety model is equally important. OpenAI's stated separation of advertising from answers means an advertiser cannot pay to change ChatGPT's response. That boundary limits a familiar form of promotional influence, but it also gives brands a clearer trust framework: the ad is visibly commercial, while the answer is intended to remain independent.
For enterprises, the immediate opportunity is not simply another media channel. It is a chance to learn how brand discovery and paid reach may coexist in conversational environments. A sensible early program should focus on:
- Context hypotheses that describe where an offer is genuinely relevant.
- Creative clarity that distinguishes a useful commercial proposition without relying on keyword insertion.
- Conversion instrumentation through the pixel or Conversions API before significant budget expansion.
- Controlled bidding tests that compare CPC and CPM objectives against internal performance goals.
- Governance review covering brand suitability, approved claims, privacy practices, and reporting ownership.
ChatGPT Ads have also expanded beyond the US over time. UK self-serve access began on June 6, 2026, while later June announcements covered activations in Japan and South Korea and further non-US expansion. Availability should therefore be checked market by market rather than assumed from the US launch.
For companies assessing AI-led discovery, paid placement is only one part of the picture. Scalevise helps teams understand how their brand and expertise appear across AI answer environments through its AI Visibility and GEO Checker. That evidence can help marketing leaders connect conversational visibility with campaign experimentation, identify gaps in their content presence, and set clearer benchmarks before allocating budget to emerging formats. Start an AI Visibility scan.
Frequently Asked Questions
When did OpenAI open ChatGPT Ads Manager to US advertisers?
OpenAI opened its self-serve Ads Manager to all US advertisers on May 5, 2026. The expansion also introduced CPC bidding alongside the existing CPM model.
How much do ChatGPT Ads cost per click?
OpenAI's Help Center recommends starting CPC bids of $3 to $5 per click. Early third-party reporting observed CPCs around $2 to $5, but actual campaign costs can change with auction conditions and inventory.
Does ChatGPT Ads support keyword or demographic targeting?
No. OpenAI uses context-based matching and ad-group context hints rather than exact-match keywords. The platform does not offer demographic targeting, though it supports geo-targeting by country or region.
Can ChatGPT Ads influence ChatGPT's answers?
No. OpenAI says ads are clearly labeled, appear separately at the bottom of conversations, and are not allowed to influence ChatGPT's answers.
What can advertisers measure in ChatGPT Ads?
Available measures include impressions, clicks, spend, click-through rate, average CPC, average CPM, and conversions measured through a pixel or the Conversions API.
Conclusion
ChatGPT Ads are now a live, expanding advertising product with self-serve US buying, CPC bidding, context-led matching, and conversion measurement options. Its most defining characteristic is not the early CPC range but the separation of ads from AI answers. Advertisers that test the channel should treat it as a developing conversational marketplace, build careful measurement from the start, and adapt their targeting and creative strategy to a model that does not depend on conventional keywords or demographics.
