OpenAI is testing ads inside ChatGPT, bringing sponsored content to the product for the first time.
The test is live for logged-in adult users in the U.S. on the free and Go subscription tiers. Plus, Pro, Business, Enterprise, and Education subscribers won’t see ads.
OpenAI announced the launch with a brief blog post confirming that the principles it outlined in January are now in effect.
OpenAI’s post also adds Education to the list of ad-free tiers, which wasn’t included in the company’s initial plans.
How The Ads Work
Ads appear at the bottom of ChatGPT responses, visually separated from the answer and labeled as sponsored.
OpenAI says it selects ads by matching advertiser submissions with the topic of your conversation, your past chats, and past interactions with ads. If someone asks about recipes, they might see an ad for a meal kit or grocery delivery service.
Advertisers don’t see users’ conversations or personal details. They receive only aggregate performance data like views and clicks.
Users can dismiss ads, see why a specific ad appeared, turn off personalization, or clear all ad-related data. OpenAI also confirmed it won’t show ads in conversations about health, mental health, or politics, and won’t serve them to accounts identified as under 18.
Free users who don’t want ads have another option. OpenAI says you can opt out of ads in the Free tier in exchange for fewer daily free messages. Go users can avoid ads by upgrading to Plus or Pro.
The Path To Today
OpenAI first announced plans to test ads on January 16, alongside the U.S. launch of ChatGPT Go at $8 per month. The company laid out five principles. They cover mission alignment, answer independence, conversation privacy, choice and control, and long-term value.
The January post was careful to frame ads as supporting access rather than driving revenue. Altman wrote on X at the time:
“It is clear to us that a lot of people want to use a lot of AI and don’t want to pay, so we are hopeful a business model like this can work.”
That framing sits alongside OpenAI’s financial reality. Altman said in November that the company is considering infrastructure commitments totaling about $1.4 trillion over eight years. He also said OpenAI expects to end 2025 with an annualized revenue run rate above $20 billion. A source told CNBC that OpenAI expects ads to account for less than half of its revenue long term.
OpenAI has confirmed a $200,000 minimum commitment for early ChatGPT ads, Adweek reported. Digiday reported media buyers were quoted about $60 per 1,000 views for sponsored placements during the initial U.S. test.
Altman’s Evolving Position
The launch represents a notable turn from Altman’s earlier public statements on advertising.
In an October 2024 fireside chat at Harvard, Altman said he “hates” ads and called the idea of combining ads with AI “uniquely unsettling,” as CNN reported. He contrasted ChatGPT’s user-aligned model with Google’s ad-driven search, saying Google’s results depended on “doing badly for the user.”
By November 2025, Altman’s position had softened. He told an interviewer he wasn’t “totally against” ads but said they would “take a lot of care to get right.” He drew a line between pay-to-rank advertising, which he said would be “catastrophic,” and transaction fees or contextual placement that doesn’t alter recommendations.
The test rolling out today follows the contextual model Altman described. Ads sit below responses and don’t affect what ChatGPT recommends. Whether that distinction holds as ad revenue grows will be the longer-term question.
Where Competitors Stand
The timing puts OpenAI’s decision in sharp contrast with its two closest rivals.
Anthropic ran a Super Bowl campaign last week centered on the tagline “Ads are coming to AI. But not to Claude.” The spots showed fictional chatbots interrupting personal conversations with sponsored pitches.
Altman called the campaign “clearly dishonest,” writing on X that OpenAI “would obviously never run ads in the way Anthropic depicts them.”
Google has also kept distance from chatbot ads. DeepMind CEO Demis Hassabis said at Davos in January that Google has no current plans for ads in Gemini, calling himself “a little bit surprised” that OpenAI moved so early. He drew a distinction between assistants, where trust is personal, and search, where Google already shows ads in AI Overviews.
That was the second time in two months that Google leadership publicly denied plans for Gemini advertising. In December, Google Ads VP Dan Taylor disputed an Adweek report claiming advertisers were told to expect Gemini ads in 2026.
The three companies are now on distinctly different paths. OpenAI is testing conversational ads at scale. Anthropic is marketing its refusal to run them. Google is running ads in AI Overviews but holding off on its standalone assistant.
Why This Matters
OpenAI says ChatGPT is used by hundreds of millions of people. CNBC reported that Altman told employees ChatGPT has about 800 million weekly users. That creates pressure to find revenue beyond subscriptions, and advertising is the proven model for monetizing free users across consumer tech.
For practitioners, today’s launch opens a new ad channel for AI platform monetization. The targeting mechanism uses conversation context rather than search keywords, which creates a different kind of intent signal. Someone asking ChatGPT for help planning a trip is further along in the decision process than someone typing a search query.
The restrictions are also worth watching. No ads near health, politics, or mental health topics means the inventory is narrower than traditional search. Combined with reported $60 CPMs and a $200K minimum, this starts as a premium play for a limited set of advertisers rather than a self-serve marketplace.
Looking Ahead
OpenAI described today’s rollout as a test to “learn, listen, and make sure we get the experience right.” No timeline was given for expanding beyond the U.S. or beyond free and Go tiers.
Separately, CNBC reported that Altman told employees in an internal Slack message that ChatGPT is “back to exceeding 10% monthly growth” and that an “updated Chat model” is expected this week.
How users respond to ads in their ChatGPT conversations will determine whether this test scales or gets pulled back. It will also test whether the distinction Altman drew in November between trust-destroying ads and acceptable contextual ones holds up in practice.
