OpenAI ads: how the ad platform works, and how to get access

Process GuideChatGPT Ads

Updated 5 min read

In short

OpenAI ads run through Ads Manager at ads.openai.com and serve as sponsored messages inside ChatGPT conversations, below the assistant’s response. As of October 2026, advertisers create campaigns in Ads Manager, set an objective, budget, and bids, then enter a relevance-weighted second-price auction that uses conversation context, ad creative, landing page, and optional context hints to decide which ad shows. Access is either self-serve in eligible countries or through OpenAI’s Ads Solutions and partner channels.

Key takeaways

  1. OpenAI’s ads platform is Ads Manager (Beta) at ads.openai.com, serving sponsored units inside ChatGPT conversations below responses.

  2. Ad selection uses a relevance-weighted second-price auction that evaluates conversation context, ad creative, landing page, and context hints.

  3. Bids are set at the ad-group level as max CPM for Reach or max CPC for Clicks, with bid-strength guidance shown in Ads Manager.

  4. Context hints at the ad-group level guide where an ad is relevant but are not exact-match keywords and do not guarantee delivery.

  5. Advertiser, creative, and placement reviews all run before serving, and access is via self-serve in eligible countries or the Ads Solutions team.

OpenAI’s ads platform is built around Ads Manager at ads.openai.com, which serves sponsored units directly inside ChatGPT conversations. This page focuses on how that system works under the hood, from auction mechanics to review layers and access paths, and you can find a separate status and eligibility overview in this ChatGPT ads guide.

What OpenAI’s ads platform is and where ads appear

Per OpenAI’s documentation as of October 2026, the ads platform is Ads Manager (Beta), available at ads.openai.com. Ads created there serve as sponsored placements inside ChatGPT conversations, positioned below the assistant’s response.

Ads Manager Beta covers campaign creation, either through a guided workflow or via bulk CSV upload. CSV upload does not support product-feed campaigns per OpenAI’s help pages. The platform also provides performance monitoring and account settings, including members, permissions, API keys, billing, and change logs.

A common misconception is that “OpenAI ads” and “ChatGPT ads” are different products. OpenAI’s help content states that Ads Manager is the platform and the inventory it sells is ads in ChatGPT, so these terms refer to the same underlying system.

How ad selection works inside a conversation

OpenAI states that ad selection uses a relevance-weighted, second-price auction that aims to maximize value for both advertisers and users. The system does not run a keyword auction in the search sense, and you do not bid on keywords.

Instead, the system evaluates expected relevance and outcomes using several signals: the context and intent of the current ChatGPT conversation, the ad’s landing page, the ad’s title and copy, and advertiser-provided context hints. When a user has ads personalization enabled, OpenAI also uses select signals from that user’s broader ChatGPT experience, per its help documentation.

Placement safeguards apply before an ad can show. OpenAI’s ad policies state that sensitive user contexts and brand-unsafe contexts are excluded from ad serving. This means even highly relevant ads are not eligible to appear in conversations that fall into those restricted contexts.

Because the auction is relevance-weighted, a higher bid does not guarantee delivery if the system predicts low relevance or poor outcomes. The matching logic is built around conversational context and policy-compliant placements rather than keyword triggers.

How the auction prices impressions and clicks

OpenAI describes its auction as a relevance-weighted, second-price system. In a second-price auction, the winning ad pays a price influenced by the next-best eligible competitor rather than its own maximum bid, adjusted for relevance in OpenAI’s implementation.

Bids are set at the ad-group level. For Reach campaigns, advertisers set a maximum CPM (cost per 1,000 impressions). For Clicks campaigns, advertisers set a maximum CPC (cost per click). OpenAI’s help content recommends a starting max CPC of $3 to $5 for Clicks campaigns.

Ads Manager provides bid-strength guidance that indicates whether a given bid is likely to be competitive or may limit delivery. This guidance is based on OpenAI’s view of the auction environment, but the exact formula for how relevance and bids interact is not published in the research set and should be treated as a black box from a practitioner perspective.

For evaluators comparing channels, the key mechanical point is that you control bids at the ad-group level and choose between impression-focused (max CPM) and click-focused (max CPC) buying, with the platform handling the relevance weighting and second-price clearing.

How context hints influence delivery

Context hints are configured at the ad-group level and are a core input to OpenAI’s matching system. Per OpenAI’s documentation, context hints describe the conversations, topics, or keywords where your product or offer may be relevant.

These hints guide matching, but they are not exact-match keywords and they do not guarantee delivery in specific conversations. OpenAI’s help pages are explicit that context hints inform the system about where an ad might be useful, while the final decision still depends on the live conversation context, policy checks, and the auction.

This corrects a common assumption that you “bid on keywords” in OpenAI ads. In practice, you set bids at the ad-group level and provide context hints as relevance signals, and the platform decides when a conversation and your ad group are a good fit.

Review layers for advertisers, creatives, and placements

OpenAI’s ad policies describe a three-layer review system covering advertisers, ad creatives and landing pages, and placements. This structure means approval is not instant in the sense of a simple account toggle.

At the advertiser level, OpenAI reviews business legitimacy through verification, account quality, and risk signals. At the creative level, each ad’s title, copy, media, and landing page are reviewed before serving. Ads or landing pages that cannot be reviewed are not eligible to run, per OpenAI’s policy documentation.

At the placement level, even approved ads only appear in conversations that comply with OpenAI’s placement policy. Placement safeguards exclude sensitive user contexts and brand-unsafe contexts, so eligibility is checked again at the moment of serving.

OpenAI states that most reviews are automated using large language models and classifiers, with human oversight. Restricted categories may require enhanced verification or manual review. This is meaningfully different from the misconception that approval is instant like a typical social ads account, because multiple automated and human-in-the-loop checks must complete before an ad can appear.

Access paths into OpenAI’s ads platform

As of October 2026, there are 2 primary access paths into OpenAI’s ads platform. The first is self-service access via Ads Manager at ads.openai.com. OpenAI’s availability documentation states that self-serve is open to businesses whose billed legal entity is in one of 63 listed countries.

The second path is through OpenAI’s Ads Solutions team and partner ecosystem. Advertisers in newly opened European markets can work with the OpenAI Ads Solutions team, agency partners, or technology partners. Larger advertisers can contact the OpenAI Ads team via an advertiser interest form, per OpenAI’s help content.

OpenAI’s documentation also notes that some new self-serve accounts can initially advertise only in their home country until identity verification approval plus a required spend threshold. This directly contradicts the misconception that self-serve is open worldwide with unrestricted geo targeting from day one.

How campaigns, ad groups, and ads are structured

OpenAI organizes accounts into campaigns, ad groups, and ads, with specific responsibilities at each level. Campaigns hold the objective, budget, and core targeting. Ad groups hold bids and context hints. Ads hold the creative assets and destination.

Per OpenAI’s help pages, campaigns contain the objective (Views, Clicks, or Conversions), budget, dates, country targeting, optional platform targeting, and optional custom audiences. A campaign’s objective cannot be changed after creation, so choosing the right objective at setup is a structural decision.

Ad groups contain the bidding model and bid amount, plus the context hints that describe where the ad group should be considered relevant. Because bids are at the ad-group level, this is where you express your willingness to pay in either max CPM or max CPC terms.

Individual ads contain the creative components: name, logo, headline, copy, landing page, and image. These elements are reviewed alongside the landing page before serving. For a step-by-step walkthrough of building this structure in the interface, see this setup guide, which focuses on execution rather than mechanics.

Correcting a related misconception, you do not manage separate “OpenAI ads” and “ChatGPT ads” accounts. The account structure described here lives entirely inside Ads Manager, and the inventory it controls is ads in ChatGPT conversations.

Frequently asked questions

OpenAI advertising runs through Ads Manager at ads.openai.com and serves sponsored units inside ChatGPT conversations, below the assistant’s response. Per OpenAI’s documentation, ad selection uses a relevance-weighted, second-price auction based on conversational context, ad creative, landing page, and context hints. This differs from traditional search ads where you typically bid on explicit keywords. In OpenAI’s system, you do not bid on keywords, and matching is driven by conversation context and policy-compliant placements.

Sources

  • This resource is for educational purposes. Ad platforms change their policies, features, and pricing frequently; verify details against the platform's current documentation before acting on them.

  • Compliance Guide

    ChatGPT ads policy: restricted categories and the rules for healthcare and finance advertisers

    A precise, dated guide to ChatGPT ad policies, including allowed, restricted, and banned categories plus standards, placements, and enforcement.

  • Process Guide

    ChatGPT ads for ecommerce: product feed campaigns, placements, and store tracking

    How ChatGPT product-feed ads work for ecommerce: formats, setup, targeting, tracking, and Shopify’s organic Catalog exposure.

  • Decision Checklist

    ChatGPT ads capabilities: the targeting, formats, and measurement available now, and what OpenAI has not built yet

    A practitioner’s matrix of ChatGPT ads targeting, formats, measurement, and controls, with documented capabilities and gaps as of October 2026.

  • Cost Guide

    ChatGPT ads cost: published pricing, CPM context from other channels, and first-test budgets

    Understand ChatGPT ads costs, budgets, and what a four-week pilot buys, using OpenAI’s published pricing and practical test math.