OpenAI Expands ChatGPT Ads With CPC Bidding and Contextual Targeting

OpenAI is expanding its ChatGPT advertising pilot beyond a CPM-only buying model, adding CPC bidding and a beta self-serve Ads Manager. The move gives advertisers a more direct way to test paid placement in ChatGPT, but it also establishes a distinctly different operating model from conventional search advertising: delivery is based on conversational context rather than keyword targeting, and the platform remains under active development.

According to OpenAI's announcement on new ways to buy ChatGPT ads, advertisers can now use CPC bidding for campaigns with a Clicks objective. OpenAI recommends beginning with maximum bids of $3 to $5 per click at the ad-group level. The company is gradually opening Ads Manager to more advertisers as part of a controlled pilot, rather than presenting the system as a fully mature, broadly available ad platform.

For marketers, the immediate story is not simply a new bid type. It is the emergence of an ad product built around the questions and tasks people bring to an AI assistant. That changes how campaigns must be planned, measured, governed, and compared with established search and display channels.

How ChatGPT ad buying works today

ChatGPT ads appear below ChatGPT responses. Instead of matching ads to exact keyword queries, OpenAI uses signals from the active conversation alongside advertiser-provided context hints to determine relevance. That makes the system contextual by design: advertisers can help describe the circumstances in which an ad should be relevant, but they are not selecting a list of exact search terms.

CPC and CPM are both supported buying approaches. CPC bidding is tied to the Clicks objective, with advertisers setting a maximum bid at the ad-group level. OpenAI's published $3 to $5 starting range gives early participants a practical test point, though it should not be read as a guaranteed market rate or a durable benchmark for every category.

Contextual matching changes campaign design

The absence of keyword targeting is a fundamental constraint, not a minor product gap. In traditional search campaigns, advertisers can build account structures around specific queries, match types, and negative keywords. ChatGPT Ads requires a different starting point: clear audience, category, use-case, and message definitions that can be represented through the context advertisers provide.

That may make the format useful for brands with a well-defined role in a broader problem-solving conversation. It may be less straightforward for advertisers whose acquisition strategy depends on tightly controlling the exact query that triggers an ad. Contextual relevance can be valuable, but it also means advertisers need to test how consistently their campaign definitions translate into appropriate conversations.

Early pricing is a signal, not a settled benchmark

Third-party reporting has placed early CPCs in the same $3 to $5 range OpenAI recommends for initial bids. The same reporting has indicated that CPMs declined from approximately $60 at launch to lower levels, with some reports citing figures around $25.

These are early observations from a pilot environment, not universal rate cards. Several factors can affect outcomes while the marketplace develops, including campaign context, advertiser demand, available inventory, creative relevance, and changes OpenAI makes to delivery. Businesses should therefore treat first campaigns as controlled learning exercises. A useful test will define a limited budget, a narrow business question, and a measurement plan before attempting to draw conclusions about channel efficiency.

Measurement, governance, and the platform opportunity

OpenAI has expanded its measurement tooling with a Conversions API and pixel-based tracking. These tools are intended to give advertisers aggregated, privacy-conscious signals that can include conversions, click-through rate, CPC, and CPM. The additions address a central operational need: paid media teams need a way to connect ad exposure and clicks to outcomes beyond the platform.

However, measurement availability is not the same as measurement maturity. Advertisers evaluating ChatGPT Ads will need to establish how conversion definitions, attribution windows, reporting consistency, and internal analytics practices fit their existing standards. The supplied research confirms the availability of the tracking approaches, but does not establish that the platform has reached parity with the reporting depth or workflow maturity of established ad ecosystems.

The pilot's controlled rollout also has practical consequences:

• Access remains gradual, as OpenAI continues to open Ads Manager to more advertisers.
• Context is the targeting foundation, requiring campaign planning that does not depend on exact keywords.
• CPC buying is now available for Clicks campaigns, alongside CPM support.
• Conversion measurement is expanding through pixel and Conversions API integrations using aggregated, privacy-conscious signals.
• Early cost reporting is fluid, so initial CPC and CPM observations should guide tests rather than set firm performance expectations.

For OpenAI, the expansion begins to clarify how an advertising business could fit within ChatGPT without turning the product into a conventional keyword-search marketplace. Ads are placed beneath responses and matched to context, which positions the unit alongside the assistant experience rather than as a replacement for the answer itself. Whether that approach delivers reliable scale, predictable performance, and advertiser confidence will depend on the next stages of the pilot.

For advertisers, governance deserves as much attention as bidding. Teams should decide which products, claims, landing pages, conversion events, and brand contexts are appropriate before launching. They should also separate exploratory spend from performance budgets until they have enough campaign data to assess quality, cost, and downstream conversion behavior.

ChatGPT advertising makes AI answer surfaces a more immediate brand-discovery channel, but paid placement does not replace understanding how an organization appears in AI-generated responses. The Scalevise AI Visibility and GEO Checker helps teams measure that baseline, identify visibility gaps, and prioritize content and entity signals before ad testing expands. This gives marketing leaders a clearer basis for connecting organic AI visibility with disciplined paid experimentation. Start an AI Visibility scan.

Frequently Asked Questions

What is new in OpenAI's ChatGPT Ads pilot?

OpenAI has expanded ChatGPT Ads beyond its CPM-only pilot model by adding CPC bidding for the Clicks objective and a beta self-serve Ads Manager that is gradually opening to more advertisers.

How does targeting work in ChatGPT Ads?

ChatGPT Ads do not use exact keyword targeting. OpenAI matches ads using signals from the current conversation and advertiser-provided context hints.

What CPC bid does OpenAI recommend for ChatGPT Ads?

OpenAI recommends that advertisers begin with maximum bids of $3 to $5 per click at the ad-group level for campaigns using the Clicks objective.

What measurement tools are available for ChatGPT Ads?

OpenAI provides pixel-based tracking and a Conversions API intended to deliver aggregated, privacy-conscious signals such as conversions, CTR, CPC, and CPM.

Conclusion

OpenAI's addition of CPC bidding and self-serve tooling makes ChatGPT Ads a more practical channel for controlled advertiser testing. Yet the platform's contextual matching model, gradual access, and evolving measurement environment mean it should be approached as an early-stage media product. The most prepared advertisers will test with clear conversion definitions, strong governance, and realistic expectations about what early pricing and performance data can show.