ChatGPT Ads: OpenAI is updating its bidding strategies
OpenAI is updating ChatGPT Ads bidding strategies by introducing Target CPA and Target CPC from 9 November 2026. These new options are gradually replacing the previous bid caps and allow for greater scope for automated optimisation. For advertisers, this change alters the way they manage acquisition costs, budgets and the performance of advertising campaigns on ChatGPT.

Key takeaways:
-
Two new bidding strategies are coming to ChatGPT Ads
: Target CPA optimises conversion campaigns around an average cost per acquisition, whilst Target CPC targets an average cost per click. -
Historical bid caps are set to be phased
out . The Max Bid and Max CPC strategies will be discontinued on 1 January 2027. Advertisers must prepare to migrate their campaigns. -
Automation is playing an increasingly important role
. OpenAI adjusts bids based on the probability of a click or conversion, the budget and the defined objectives. -
Data quality is becoming a key
factor. The signals transmitted by the OpenAI Pixel or the Conversions API influence the system’s ability to optimise ad delivery.
In summary, ChatGPT Ads is moving towards results-based management rather than manual bid control. Advertisers must now place as much importance on the reliability of tracking as on defining their profitability objectives.
ChatGPT Ads: three bidding strategies for managing campaigns
With this update, OpenAI is offering three approaches to managing ad auctions in ChatGPT Ads. They serve different purposes, but all rely on automated delivery optimisation.
Intelligent Bidding: maximising results within a given budget
Formerly known as Maximize Results (Autobid), the Intelligent Bidding strategy automatically adjusts bids and ad delivery to achieve the desired results. The advertiser sets their budget and business objective, without specifying a target average cost. The system then handles individual bidding decisions.
Target CPA: optimising the cost per conversion
Available from 9 November 2026 for new conversion campaigns, Target CPA allows you to set a desired average cost per acquisition. For example, with a Target CPA of $50, OpenAI aims to achieve conversions around this average. Some may cost more, others less.
Target CPA is an optimisation target, not a cost guarantee. The actual average CPA achieved may exceed the defined target.
Target CPC: managing the average cost per click
Target CPC follows a similar approach for campaigns aimed at generating clicks. Unlike the old Max CPC, which capped individual bids, this strategy allows the system to bid above or below the target depending on the opportunities available.
The aim is to maximise clicks within the available budget, whilst ensuring that the average CPC is in line with the target.
| Strategy | Main objective | Control mode |
| Intelligent Bidding | Maximising results | Budget, with no target cost |
| Target CPA | Generate conversions | Average cost per conversion |
| Target CPC | Generate clicks | Average cost per click |
What the new ChatGPT Ads auctions mean for advertisers
The main change involves a shift from monitoring individual bids to a target average cost approach.
OpenAI states that advertising bids are still determined by an auction system. The final amount of the winning bid determines the price paid.
It is also important to distinguish between the optimisation strategy and the billing model. A campaign optimised for conversions may be billed on a per-click or per-impression basis, depending on its settings. This distinction is important for advertisers who are used to SEA management models.
Another change: the old Max Bid and Max CPC strategies will be phased out on 1 January 2027. Existing campaigns will need to switch to the new options.
For API integrations, OpenAI is also planning to introduce new ‘Target CPA’ and ‘Target CPC’ fields, and a transition will need to be prepared before this deadline.
Opportunities and risks: automation depends on signal quality
Changes to ChatGPT Ads’ bidding strategies offer greater flexibility in optimising performance.
In particular, the system can adjust its bids based on the estimated probability of receiving a click or a conversion. This approach reduces the need to manage each bid manually.
However, it raises several points to bear in mind. An overly restrictive target may limit reach. OpenAI notes that a target set below recently observed performance levels may prevent the campaign from spending its entire budget.
The quality of tracking is another key challenge. For conversion campaigns, OpenAI relies on events transmitted via its Pixel, its Conversions API, or both.
The Event Quality Score (EQS) is used to assess the usefulness of the information received for measurement purposes. However, a high score does not guarantee a sufficient volume of conversions to establish a scale.
Advertisers must therefore monitor the quality of events, their volume and the results actually generated, all at the same time.
How can you adapt your ChatGPT Ads strategy?
The transition to new bidding strategies requires a step-by-step approach.
Three actions must be given priority:
- Analyse historical performance. Compare the actual CPA or CPC with the campaign’s financial targets, over a consistent period.
- Check the reliability of conversions. Monitor event reception, potential duplicates, delays and quality alerts in Ads Manager.
- Set realistic targets. Gradually adjust the targets based on actual expenditure, results and average costs.
OpenAI gives an example: a campaign that has spent $400 on 10 attributed purchases has a CPA of $40. Setting a target of $20 straight away does not mean that the system will be able to halve that cost.
The challenge, therefore, is to strike a balance between economic ambition and actual distribution capacity.
Expert advice: managing ChatGPT Ads based on business value
The introduction of Target CPA and Target CPC brings the management of ChatGPT Ads closer to the automation approaches already used in paid search.
For advertisers, the priority should not be to systematically seek out the lowest target cost.
A campaign may show an attractive CPA without necessarily generating the most profitable customers. Conversely, a higher customer acquisition cost may be acceptable if the business value gained justifies it.
It is therefore useful to compare advertising metrics with business data: lead conversion rates, order value, profit margins and the quality of customers acquired.
The key performance indicators remain CPA or actual CPC, the volume of results, budget spent and the quality of conversion signals.
This approach helps to avoid optimisation that is limited solely to media costs.
Conclusion: a new step forward in the automation of ChatGPT Ads
With the new ChatGPT Ads bidding strategies, OpenAI is strengthening its approach to automated management based on advertisers’ objectives.
Target CPA and Target CPC offer greater flexibility, but require realistic targets and reliable conversion data.
The first step is to audit existing campaigns before the old bid controls are phased out, which is scheduled for 1 January 2027.
Our agency, GEA, supports you in analysing new advertising channels, structuring your acquisition strategies and managing performance across search and AI platforms.
FAQ – ChatGPT Ads Bidding Strategies
Target CPA aims for an average cost per conversion, whilst Target CPC aims for an average cost per click. Both strategies allow OpenAI to automatically adjust bids according to the defined objectives. However, they do not guarantee a fixed price: the actual average costs incurred may be higher or lower than the targets.
Intelligent Bidding automatically adjusts bids and ad delivery to achieve the desired results within the available budget. Unlike the Target CPA and Target CPC strategies, this option does not require you to set a target average cost. It is suitable for advertisers who wish to give the optimisation system greater flexibility.
The Max Bid and Max CPC strategies will be phased out on 1 January 2027 for existing campaigns. OpenAI is allowing advertisers to begin migrating to Target CPA and Target CPC from 9 November 2026. Partners using the API must also adapt their integrations to the new fields before the old settings are removed.
No, Target CPA does not guarantee a maximum cost per conversion. It sets an average that the system aims to achieve by adjusting its bids. Some conversions may cost more than the target. The observed average CPA may also exceed it, particularly depending on available opportunities and the quality of the signals provided.
A ChatGPT Ads campaign may under-spend if its cost targets are too restrictive, if advertising opportunities are limited, or if conversion signals are insufficient. OpenAI recommends checking recent performance, event reception and quality diagnostics in Ads Manager before adjusting targets.