ChatGPT Ads are labeled sponsored placements that OpenAI sells to businesses and shows to users on ChatGPT's free tiers. They are self-serve in many countries and cheap enough to test, but OpenAI has published no performance benchmarks. For most brands, the sensible move is a small, tracked test rather than a large commitment.
This guide covers how the ads work, which countries can buy them, what they cost, how to measure results, and how they compare with Google and Meta Ads. It ends with a Test / Wait / Skip framework and a 30-day test plan. Platform details change often, so each fact names its source type and was checked on the date above.
What Are ChatGPT Ads? Quick Answer and Verdict
Definition: ChatGPT Ads are labeled sponsored placements that OpenAI sells through its Ads Manager platform and shows to signed-in users on ChatGPT's Free and Go plans. Targeting runs on conversation context rather than keywords or interest audiences, with no PII-based targeting or retargeting. Ads appear separately from ChatGPT's answers and do not change what an answer says.
OpenAI's US pilot began on February 9, 2026, for logged-in adults on the Free and Go plans. Self-serve Ads Manager access opened on May 5, 2026. On October 21, 2026, OpenAI said ads are available in 89 countries and that more than 100,000 advertisers have run campaigns.
Not every ChatGPT user sees ads. Plus, Pro, Business, Enterprise, and Edu accounts remain ad-free. That means the audience is the free and low-cost segment of ChatGPT users, not everyone.
The verdict in three lines:
Test: if you have working conversion tracking, a margin that can absorb a $3 to $5 click, and a budget you can afford to lose.
Wait: if your margins are thin, your tracking is unreliable, or one lost month of budget would hurt.
Skip: if your category is not eligible, or you need city-level targeting.
One mistake people often make is treating the advertiser count as proof of profit. A count of 100,000 advertisers measures participation, not return on ad spend.
How Do ChatGPT Ads Work?
ChatGPT Ads work by matching an advertiser's ad to the topic of a user's conversation. The advertiser sets context hints and a bid, an auction decides which ad wins, and a labeled sponsored unit appears next to the answer. The advertiser pays on impressions or clicks.
The sequence, from a user's question to a recorded sale:
A user asks ChatGPT something. The conversation topic is the targeting signal.
The advertiser has set context hints. These hints guide ad matching, but they are not exact-match keywords and do not guarantee delivery in specific conversations.
An auction runs. OpenAI uses a relevance-weighted, second-price auction, so the highest bid alone does not win.
A sponsored unit appears. It is labeled and separate from the answer.
The user clicks. The click carries a reference that later ties a conversion back to the ad.
The landing page records an outcome. A purchase, lead, or sign-up is sent back to Ads Manager through tracking.
On privacy, OpenAI says ads are labeled, do not influence answers, and do not give advertisers access to users' conversations. Per OpenAI's description, advertisers buy topical context, not individual people.
A better way to think about targeting is in conversation topics, not keywords. A keyword plan asks "what will people type?" A topic plan asks "what will people be trying to work out when they would want my product?" Some examples:
Ecommerce: gift ideas for a new homeowner, or choosing running shoes for flat feet.
SaaS: comparing invoicing tools for freelancers.
Travel: planning a family trip with a toddler.
Keyword-style thinking is the most common early error, and it produces hints that are too narrow to match anything.
Where Are ChatGPT Ads Available? (US, UK, Canada, Australia, Bangladesh, UAE)
ChatGPT Ads are sold in 89 countries as of October 21, 2026, according to OpenAI. OpenAI's help center lists Australia, Canada, the United Arab Emirates, the United Kingdom, and the United States as self-service available. Bangladesh was named in the October expansion. Availability changes often, so confirm your market on OpenAI's Ads Manager Availability page.
Country | Status | Buy route | Notes |
United States | Available | Self-serve Ads Manager | Launch market, with the most mature setup |
United Kingdom | Available | Self-serve (listed) | Confirm consent setup for your tracking |
Canada | Available | Self-serve (listed) | Confirm currency and billing |
Australia | Available | Self-serve (listed) | Confirm currency and billing |
United Arab Emirates | Available | Self-serve (listed) | Check whether Arabic prompts are eligible |
Bangladesh | Named in Oct 21 expansion | Check your account | Advertisers in new countries can use technology partners or self-service Ads Manager |
Verified October 2026 from OpenAI's help center listing and OpenAI's October 21 announcement. The live page is the authority.
A common issue is assuming that "available in your country" means "ready for your business." Before signing up, confirm three things:
Billing: which currencies and payment methods your account accepts.
Language: whether prompts in your customers' language, such as Bangla or Arabic, can trigger ads. Confirm this in OpenAI's documentation rather than assuming.
Route: whether your country is self-serve or partner-only.
Targeting granularity matters locally. Ads Manager Beta currently targets at the country level, so you select countries rather than cities or regions. A Dhaka or Dubai business can choose Bangladesh or the UAE, but not a city. Some sources describe newer geotargeting options, so check your own account before planning around either.
ChatGPT Ads Formats, Targeting and Campaign Setup
A ChatGPT ad campaign is built in OpenAI's Ads Manager in four decisions: objective, budget, country and context hints, plus the creative and landing page. Campaigns follow a three-level structure, and the objective you choose sets your pricing model. Most brands should start with one campaign and one objective.
Confirm availability and create an account at ads.openai.com.
Add billing details.
Choose an objective. Views optimizes for reach and bills on CPM. Clicks optimizes for traffic and bills on CPC.
Set a budget. Since June 9, 2026, daily budgets work as averages, and OpenAI can spend up to twice your daily amount on a busy day. You can change the amount later but cannot switch the budget type.
Select countries and write context hints.
Create the ad. Write for someone mid-conversation, with a clear promise that matches the landing page.
Check the landing page and tracking. Test the page on mobile, and fire a test conversion before spending.
Launch and review.
Category rules apply. One industry guide lists eligible categories as consumer products, local services, travel, entertainment, digital products and education, with more verticals expected as compliance systems develop. Treat that as a starting point and check OpenAI's advertising policies for your category.
Minimum spend conflict. One source says there is no minimum spend. Another says the current minimum is $25 per day per campaign. Rules like this change, so confirm the number inside your own Ads Manager before budgeting.
Pre-launch checklist:
The landing page loads fast on mobile.
The page answers the same question the ad implies.
A test conversion appears in your reporting.
Your UTM parameters are set.
The ad claims match your policies and your category's rules.
Testing the tracking before the first click is the step teams skip most often, and it usually costs them their first week of data.
ChatGPT Ads Cost and Pricing
ChatGPT Ads are priced by auction on either CPM (cost per thousand impressions) or CPC (cost per click). Reported CPMs have fallen from about $60 toward $25, and OpenAI suggests starting CPC campaigns with a $3 to $5 maximum bid. The final price depends on competition.
Model | How billed | Reported range | Best used for |
CPM (Views) | Per 1,000 impressions | Reported drop from $60 at launch to as low as $25 | Awareness and reach tests |
CPC (Clicks) | Per click | OpenAI suggests a $3 to $5 starting max bid | Traffic and conversion tests |
Ranges are reported figures and may change as the auction matures.
A bid is a ceiling, not a price. For CPM campaigns, the default max bid has been around $60, but these are ceilings set by the advertiser, not fixed prices. The second-price auction usually means you pay less than your maximum.
Google keyword costs are no guide here. Keyword tools show Google Ads bids for searches about ChatGPT ads, which says nothing about what ChatGPT's own auction will charge you. The practical advice is to run a small test, read your own average CPC after the first week, and plan from that number.
The number that matters is break-even. The formula:
Break-even conversion rate = average CPC ÷ allowable CPA
Hypothetical illustration, not a benchmark: if your average CPC is $4 and you can afford to pay 80towinacustomer,youneedabout5%ofclickstoconvert(4 ÷ $80). If your site converts closer to 2%, the channel loses money at that CPC no matter how good the click-through rate looks.
Teams often run into trouble when they budget by impressions and never calculate break-even.
ChatGPT Ads Conversion Tracking and Measurement
Tracking ChatGPT Ads conversions takes four steps: install the OpenAI Pixel, add the Conversions API, deduplicate events with a shared event ID, and verify the results in GA4 or your store's own data. Ads Manager reports clicks and conversions, but your own revenue data is the ground truth.
Setup checklist:
Install the OpenAI Pixel on your landing and confirmation pages.
Add the Conversions API for server-side events.
Send the same event ID through both so each conversion counts once.
Tag every ad URL with UTM parameters, such as a utm_source value that identifies ChatGPT.
Confirm a test purchase or lead appears in Ads Manager and in GA4.
Here is how the pieces connect. When someone clicks an ad, OpenAI appends a click reference called oppref to the landing page URL. The Pixel captures it in a first-party cookie, the Conversions API can send it with server-side events, and a shared event ID deduplicates the same conversion.
The two methods cover different gaps. The Conversions API sends events server-to-server, so ad blockers and cookie restrictions do not interrupt it. Running both requires deduplication to avoid double counting. This usually breaks when only one method is installed, or when the two send different event IDs.
On Shopify or WooCommerce, check whether a native integration or a tag manager route exists for your setup. If not, add the base code and event code manually, then test.
For quality checks and outside validation, OpenAI named its first third-party measurement partners on October 5, covering attribution, incrementality and brand measurement, and pointed advertisers to an Event Quality Score for judging conversion signals.
What Ads Manager reports | What you must measure yourself |
Impressions, clicks, spend, CTR, average CPC, average CPM and conversions | Revenue and margin per order |
Attributed conversion events | Lead quality and close rate |
Spend by campaign, ad group and ad | Customer acquisition cost over time |
Click and conversion counts | Assisted and repeat purchases |
What matters most here is trusting your own system first. If Ads Manager and your store disagree, reconcile against actual revenue before changing a bid.
How Do ChatGPT Ads Perform? What the Data Actually Shows
No official performance benchmark exists for ChatGPT Ads. OpenAI has published no cross-advertiser benchmark for CTR, conversion rate, CPA or ROAS. Published figures come from analysts, agencies and vendors, and they vary widely by business, category and conversation type.
Claim | Source | Sample | Confidence |
No official benchmark exists | Independent review of OpenAI documentation | n/a | High |
CTR around 0.68% | Similarweb analysis, May 2026 | Not stated | Medium |
CTR 0.91% vs 6.4% on Google Search | Adthena data as reported by a third party | Early pilot weeks | Medium |
1.49x ROAS | Opascope public case study, 15 days, $60,000 spend | One advertiser | Low to medium |
Daily ROAS swings from 0.2x to 2.9x | A 2026 benchmark roundup | Not stated | Low |
Conversion rates 2 to 3x higher than Google Shopping | Marketing agency article | Not stated | Low (advertiser claim) |
CTR about 3x and conversions near 2x vs comparable formats | Criteo update, vendor-reported | Vendor data | Low to medium |
A test was suspended over concerns about click data | 36Kr report on one advertiser | One test | Low |
The pattern across sources is lower click-through, with higher conversion claimed on the clicks that do happen. A lower CTR is partly a placement effect, because the ad appears after the model has already answered. That is why CTR alone is the wrong judge. A channel can have a low CTR and still be profitable, or a decent CTR and still lose money.
Spend levels show demand but not profit. ChatGPT ads crossed a $1 billion annualized run rate in late August 2026, fewer than 200 days after launch. Revenue to OpenAI shows advertisers are buying, not that they are profitable.
ChatGPT Ads vs Google Ads vs Meta Ads
ChatGPT Ads target conversational context, Google Ads target search queries, and Meta Ads target audiences and feed behavior. ChatGPT Ads are the youngest and least measurable of the three, which makes them better for testing than for core budget.
Criteria | ChatGPT Ads | Google Ads | Meta Ads |
Intent signal | Conversation topic, often early research or comparison | Search query, often high purchase intent | Interest and behavior, often discovery |
Targeting | Context hints, country level | Keywords, audiences, location | Audiences, interests, lookalikes, location |
Bidding | CPC or CPM auction | Mostly CPC, with automated options | Mostly CPM-based with optimization goals |
Measurement maturity | New: Pixel, Conversions API, Event Quality Score | Mature, widely integrated | Mature, widely integrated |
Minimum spend | Reported conflict; check your account | Set by you | Set by you |
Best use | Early-funnel research and comparison moments | Capturing existing demand | Creating and retargeting demand |
Biggest limitation | Thin benchmarks and reporting | Rising competition and costs | Creative fatigue and privacy changes |
The practical difference is where each channel sits in the funnel. Google captures people who already know what they want. Meta interrupts people who have not started looking. ChatGPT reaches people while they are working out what to buy, which is a different moment and often a different cost.
On click-through alone, ChatGPT trails Search. Adthena data reported about 0.91% for ChatGPT against 6.4% for Google Search. One mistake people often make is replacing Google budget with ChatGPT budget on that basis. The two channels answer different questions, so a test should add budget rather than move it.
Paid ChatGPT Ads vs Organic ChatGPT Visibility (AEO/GEO)
Paid ChatGPT Ads do not change what ChatGPT recommends or cites. OpenAI says ads do not influence the assistant's answers. A brand that wants to appear inside an answer needs organic Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO), which means clear entity signals, citable content, and consistent brand information across the web.
Paid and organic AI visibility are separate workstreams that can reinforce each other. Ads buy a labeled placement now, while organic work builds the authority AI systems draw on over time. Riha Web Tech treats them as separate budgets with separate goals.
Should Your Brand Use ChatGPT Ads Yet? Test, Wait or Skip
Are ChatGPT Ads worth it? For many businesses, a small tracked test is worth it. No official benchmark exists, so the return has to come from your own data. Brands with working tracking and healthy margins should test, and brands with thin margins or unreliable tracking should wait.
Business type | Verdict | Condition | First test |
Small business | Test | Tracking works, budget is genuinely expendable | One CPC campaign, 2 to 3 conversation themes |
Ecommerce | Test | Gross margin comfortably covers a $3 to $5 click | Product-research themes tied to top sellers |
SaaS and B2B | Test cautiously | Deal size supports the CPC, and leads reach a CRM | Comparison and "how to choose" themes |
Startup | Wait | No conversion data or runway to absorb tests | Build tracking first, then revisit |
Local services | Wait or skip | Targeting is country-level, so city-only businesses pay for unusable reach | Only if you serve a whole country |
Three criteria decide most cases:
Margin: can one click at the reported range still leave a profit at your conversion rate?
Deal size: high-value sales absorb expensive clicks better than low-value ones.
Eligibility: does your category fit OpenAI's advertising policies?
Skip when your category is ineligible, when your product sells only in one city, or when you cannot track conversions at all. Few guides say when not to buy, but that is often the more useful advice. A business that cannot measure the channel cannot judge it, however good the platform becomes.
Limitations and Risks of ChatGPT Ads
ChatGPT Ads have four main limits: a beta platform, volatile pricing, thin reporting, and an audience limited to free tiers. Each one affects how much budget a brand can reasonably defend.
Beta status: Some reporting features remain limited, and the $3 to $5 CPC range will likely shift as more advertisers enter.
Volatile pricing: CPMs have already moved from about $60 to as low as $25 in months.
Thin reporting: Ads Manager shows delivery and conversions, but not the conversation context behind a click.
Audience skew: ads appear only to Free and Go users, not to paid-plan users.
Brand suitability: you buy topics, not specific conversations, and delivery is not guaranteed.
Privacy and regional rules: consent and data requirements differ by country, so check them before tagging pages.
When sources disagree. The minimum-spend rules and targeting options are reported differently across sources. A good way to handle this is to treat OpenAI's help center as the authority, treat third-party pages as leads to verify, and confirm any number that affects your budget inside your own Ads Manager. Note the date you checked it.
A 30-Day ChatGPT Ads Test Plan (and Common Mistakes)
A good ChatGPT Ads test has four steps: set up and verify tracking, launch one campaign on one objective, judge results against break-even math at day 14, and decide whether to scale, adjust or stop at day 30.
Week 1: Setup and tracking.
Install the Pixel and Conversions API.
Confirm a test conversion appears in Ads Manager and your own analytics.
Set one Clicks campaign, one country, and 2 to 3 conversion themes.
Decide your allowable CPA and break-even conversion rate before launch.
Week 2: Launch and observe.
Avoid edits for the first few days, so the auction has data to work with.
Check delivery, CPC, and the first conversions. At day 14, compare actual CPC and conversion rate to your break-even numbers.
Week 3: Adjust.
Pause themes that spend without converting.
Test a second creative on the best theme.
Week 4: Decide.
At day 30, judge on cost per conversion and revenue, not CTR.
Scale only if actual cost per conversion is below your allowable CPA on enough conversions to trust.
Budget sizing is arithmetic. If a $25 daily minimum applies to your account, a 30-day single-campaign test is about $750. At a hypothetical $4 CPC, that buys roughly 187 clicks. At a 2% conversion rate, that is only about 4 conversions, which is a thin sample. Decide in advance how many conversions you need before a result counts.
Common mistake | Better approach |
Judging by CTR alone | Judge by cost per conversion and revenue |
Launching without tracking | Test the Pixel and API before spending |
Using keyword-style hints | Write hints around conversation topics |
Scaling after a lucky first week | Wait for enough conversions at your break-even |
Counting conversions twice | Send one shared event ID through both methods |
Using Google CPC as a budget guide | Plan from your own first-week average CPC |
Conclusion: Should You Test ChatGPT Ads?
ChatGPT Ads are a legitimate new channel, but not yet a proven one. Three points matter most:
They are worth a small, tracked test for brands with working tracking and healthy margins.
Profit depends on your own numbers, because no official benchmark exists. Calculate break-even before you spend.
Paid ads and organic AI visibility are separate. Ads do not change what ChatGPT recommends.
Your next step is to run the 30-day plan with one campaign and one objective, and write down your stop and scale rules before launch. If you want help setting up tracking or structuring a test, see Framecipher’s paid advertising services.

