What do ChatGPT ads cost? The short answer
Here is the direct answer most people are searching for: as of 2026, ChatGPT ads cost between $25 and $60 per thousand impressions on the CPM model, or between $3 and $5 per click on the newer CPC model. Premium verticals like finance and technology can push CPMs toward $65. The self-serve platform has a realistic entry point of around $5,000 per month, with daily spend floors near $500.
Those ranges matter because ChatGPT is no longer a niche experiment. OpenAI reports more than 800 million weekly ChatGPT users, and ads are shown to the logged-in Free and Go ($8/month) tiers in the United States, with expansion to Canada, Australia, and New Zealand underway in 2026. That is an enormous, high-intent audience that did not exist as an ad channel two years ago, which is exactly why advertisers want to know what it costs before competitors bid the prices up.
The broader context is a market rushing toward AI-driven advertising. The AI in advertising market is projected to grow from about $14.1 billion in 2026 to $36.3 billion by 2030, expanding at roughly 27% per year. ChatGPT ads sit at the center of that shift, and early pricing reflects both scarcity of inventory and strong demand.
$25 to $60
Typical CPM range for ChatGPT ads in 2026, reaching ~$65 in premium verticals
The rest of this guide breaks those numbers down: how the pricing models work, how cost varies by industry, what a minimum budget actually buys, and a step-by-step calculator to estimate impressions, clicks, conversions, and cost per acquisition before you commit any spend.
The three pricing models: CPM, CPC, and CPA
ChatGPT ads run on a relevance-weighted, second-price auction, and OpenAI now supports two live pricing models with a third in development. Understanding which one you are billed under is the foundation of any accurate cost estimate.
CPM: pay per thousand impressions
CPM (cost per mille) was the original model when ChatGPT ads launched in February 2026. You pay a fixed rate every time your ad is shown 1,000 times, whether or not anyone clicks. Rates run from $25 to $60, with most advertisers landing between $35 and $50 depending on how competitive their topic clusters are. CPM rewards high click-through rates: the more clicks you earn per thousand impressions, the lower your effective cost per click.
CPC: pay per click
CPC (cost per click) was added in April 2026 for performance advertisers who wanted to pay only for engagement. Starting bids range from $3 to $5, and you pay nothing for impressions that do not convert into clicks. CPC makes budgeting predictable and caps your downside, which is why most first-time advertisers should start here.
$3 to $5
Starting CPC bid range on ChatGPT ads, added April 2026
CPA: coming later in 2026
Cost-per-acquisition bidding, where you set a target cost per conversion and the algorithm optimizes delivery to hit it, is in development and expected later in 2026. Until it launches, you model your CPA manually using the calculator further down this page.
| Model | You pay for | Typical range | Availability |
|---|---|---|---|
| CPM | Every 1,000 impressions | $25 to $60 | Live since February 2026 |
| CPC | Every click | $3 to $5 | Added April 2026 |
| CPA | Every conversion | Target-based | In development, later 2026 |
Which model is cheaper depends entirely on your click-through rate. For a full break-even analysis of when CPM beats CPC (and the exact CTR threshold where the math flips), read our CPC vs CPM bidding strategy guide. For this cost guide, the key takeaway is simple: CPC gives you predictable spend, and CPM gives you cheaper clicks once your CTR climbs above roughly 1%.
ChatGPT ad costs by industry
ChatGPT ad costs are not uniform. Because the auction is relevance-weighted and topic-driven, CPMs rise in verticals where advertisers compete hardest for the same conversations. The pattern mirrors what we see on every other platform: high-value B2B and regulated industries pay a premium, while lifestyle and entertainment categories sit at the lower end.
For reference, the same dynamic plays out on Google, where the 2026 average cost per click is $5.42 across all industries but climbs to $9.87 for legal services and drops to $1.63 for arts and entertainment. ChatGPT CPMs follow a similar spread. The table below shows estimated CPM ranges by vertical based on early advertiser data.
| Industry / vertical | Estimated ChatGPT CPM | Relative competition |
|---|---|---|
| Technology & SaaS | $45 to $65 | Very high |
| Finance & insurance | $45 to $65 | Very high |
| Legal & professional services | $40 to $60 | High |
| B2B software & services | $38 to $55 | High |
| Healthcare & wellness | $35 to $52 | Moderate to high |
| Education & e-learning | $30 to $45 | Moderate |
| E-commerce & retail | $30 to $45 | Moderate |
| Travel & hospitality | $28 to $42 | Moderate |
| Lifestyle, arts & entertainment | $25 to $38 | Lower |
Two caveats. First, sensitive topics such as health conditions, mental health, and politics are excluded from ad serving entirely, so a wellness advertiser will only reach non-sensitive conversations. Second, these are estimates from a young platform; your actual CPM depends on your relevance score, which is why creative quality (covered below) is such a powerful lever on cost.
Minimum budget: what it takes to start
When OpenAI ran its managed pilot, the entry point was a reported $200,000 minimum commitment, which locked out everyone except large brands. That changed in April 2026 when the self-serve platform launched at ads.openai.com and dropped the managed-pilot minimum. The practical floor is now roughly $5,000 per month, with daily spend floors around $500.
~$5,000/mo
Practical self-serve minimum, with $500 daily spend floors
What a small budget realistically buys
At a blended $45 CPM, a $5,000 monthly budget buys roughly 111,000 impressions. If your ad earns a 1% click-through rate, that is about 1,111 clicks. If your landing page converts 3% of those clicks, you get around 33 new customers or leads for the month. That is a meaningful test, but it is not enough volume to optimize a dozen topic clusters at once.
The single biggest mistake at the minimum budget is spreading spend too thin. A $5,000 budget split across ten topic clusters gives each cluster only 11,000 impressions, far too few to gather statistically useful data. Concentrate your first month on one or two of your highest-intent clusters, prove the funnel, then expand. Starting focused is the difference between learning something and burning $5,000 on noise.
If $5,000 per month is out of reach today, the smartest move is to prepare your creative library now so you can launch efficiently the moment your account is live. That is where an AI generator pays for itself: you can build and test dozens of ChatGPT-format variations for the price of a single agency creative, then deploy the winners with zero production lag.
The ChatGPT ads cost calculator
You do not need a spreadsheet to estimate what a ChatGPT campaign will cost. You need five inputs and five steps. Work through them in order and you will have a defensible estimate of impressions, clicks, conversions, and cost per acquisition before you spend anything.
The five-step method
- Start with your monthly budget. This is the number you are willing to commit.
- Estimate impressions. Divide your budget by your CPM, then multiply by 1,000. (Under CPC, skip to step 3 and divide budget by your CPC bid to get clicks directly.)
- Estimate clicks. Multiply impressions by your expected click-through rate (CTR).
- Estimate conversions. Multiply clicks by your landing-page conversion rate (CVR).
- Calculate cost per acquisition. Divide your budget by conversions.
A worked example
Assume a blended $45 CPM, a 1% CTR, and a 3% landing-page conversion rate. A $5,000 budget produces about 111,000 impressions, which yields roughly 1,111 clicks (a $4.50 effective cost per click), which yields about 33 conversions, for a cost per acquisition near $150. The table below scales that same math across three common budget levels.
| Monthly budget | Impressions (at $45 CPM) | Clicks (1% CTR) | Conversions (3% CVR) | Cost per acquisition |
|---|---|---|---|---|
| $2,500 | ~55,600 | ~556 | ~17 | ~$150 |
| $5,000 | ~111,100 | ~1,111 | ~33 | ~$150 |
| $10,000 | ~222,200 | ~2,222 | ~67 | ~$150 |
~$150 CPA
Estimated cost per acquisition at $45 CPM, 1% CTR, and 3% conversion rate
Notice that CPA stays flat as budget scales. That is the math working as intended: doubling spend doubles conversions when your rates hold. The two inputs that move your CPA the most are CTR and CVR, not budget. Improving your CTR from 1% to 1.5% cuts your effective cost per click by a third. Improving your conversion rate from 3% to 4.5% cuts your CPA by a third again. This is why creative and landing-page quality are the highest-leverage cost controls you have.
The true cost per acquisition
Comparing ChatGPT to other platforms on CPM or CPC alone is misleading, because the click cost tells you nothing about what happens after the click. The metric that matters to your business is cost per acquisition, and that is where ChatGPT’s higher impression price is offset by dramatically higher conversion quality.
Users who click a ChatGPT ad have already self-qualified. They described their problem in the conversation, specified their constraints, and are actively evaluating options when your ad appears. That is a fundamentally warmer click than a broad-audience Meta impression or even a Google search click. Early advertiser data shows ChatGPT ads converting 1.5 to 4× better than Google Search in matching verticals.
1.5 to 4×
Higher conversion rate for ChatGPT ads vs Google Search in matching verticals
| Platform | Typical CPC | Typical CPM | Click-to-conversion | Effective CPA |
|---|---|---|---|---|
| ChatGPT (CPC) | $3 to $5 | Click-based | 2.0 to 4.0% | $75 to $250 |
| ChatGPT (CPM) | ~$4.50 effective | $25 to $60 | 2.0 to 4.0% | $110 to $225 |
| Google Search | ~$5.42 avg | Search-based | 1.5 to 3.0% | $180 to $360 |
| Meta Ads | $0.70 to $1.90 | $11 to $14 | 0.8 to 1.6% | $45 to $190 |
| LinkedIn Ads | $5 to $12 | $30 to $80 | 1.5 to 2.5% | $200 to $800 |
Read this table carefully. Meta looks cheapest on a per-click basis, but that low cost reflects broad, lower-intent reach: you buy many cheap clicks that convert at low rates. Google’s $5.42 average CPC combined with typical conversion rates produces an effective CPA of $180 to $360. ChatGPT’s higher CPM is repaid by its conversion advantage, landing effective CPA in a competitive $75 to $250 band despite the premium impression price. When you buy intent instead of reach, the cost per click is the wrong scoreboard.
To measure this correctly for your own campaigns, set up conversion tracking from day one. Our ChatGPT ads ROI measurement guide walks through attribution, UTM structure, and how to calculate true return on ad spend.
The hidden cost of creative production
Every cost breakdown so far has been about media spend, the money that goes to OpenAI. But there is a second budget line that most advertisers underestimate and that often exceeds media cost for smaller campaigns: creative production. You cannot run ChatGPT ads without headlines, descriptions, and images, and you cannot optimize without producing many variations to test.
Traditional production is slow and expensive. An agency charges $500 to $3,000 per creative concept and takes one to three weeks. A freelancer is cheaper but still $150 to $800 per creative and several days of back-and-forth. Even an in-house designer carries a loaded cost of $100 to $300 per creative once you account for salary and time. None of these scale to the dozens of variations that effective testing requires.
| Production method | Cost per creative | Turnaround | Best for |
|---|---|---|---|
| Traditional ad agency | $500 to $3,000 | 1 to 3 weeks | Big-budget brand campaigns |
| Freelance designer | $150 to $800 | 3 to 7 days | One-off creatives |
| In-house designer | $100 to $300 (loaded) | 1 to 5 days | Ongoing brand work |
| Stock assets + DIY editor | $20 to $100 | Hours to days | Tight budgets, manual effort |
| Lapis | About $4 | Under 3 minutes | Endless on-brand variations, every platform |
About $4
Lapis cost per production-ready creative vs $500 to $3,000 at an agency
Lapis generates production-ready ads for Meta, Google, Reddit, LinkedIn, and ChatGPT from a single text prompt in under three minutes, at roughly $4 per creative. It is the first and only AI ad platform ready for ChatGPT advertising, and it outputs headlines within the 50-character limit, descriptions within the 100-character limit, and images sized for ChatGPT’s thumbnail format. Its Brand Intelligence crawls your website automatically so every creative is on-brand, and it can pull products directly from Shopify or Amazon with @product mentions. For a full breakdown of what AI ad tools charge, see our AI ad generator pricing comparison.
Do not forget the other hidden costs either: dedicated landing pages for each intent cluster, and measurement infrastructure to track conversions. Lapis includes built-in Web Analytics and conversion tracking, so the measurement piece is handled without stitching together separate tools.
How to lower your ChatGPT ad costs
Because the ChatGPT auction is relevance-weighted, the most effective way to lower cost is not to bid less; it is to raise your relevance score. A high-relevance ad wins the same auctions at a lower price, so quality and cost are directly linked. Here are the levers that move your costs down.
Improve creative relevance
Precise context hints and conversational copy raise your relevance score, which lowers both your effective CPM and your effective CPC through the auction mechanics. A specific ad (“CRM for service teams under 20 people”) beats a generic one (“business software”) at a lower bid. Testing many creative angles is how you find the high-relevance winners.
Forecast before you spend
The most expensive mistake is spending real budget to discover that a creative underperforms. Lapis performance forecasting predicts impressions, clicks, CTR, and leads for each creative before launch, so you can kill weak variations on paper instead of paying to learn the same lesson. Combined with the calculator above, forecasting turns your CPA estimate from a guess into a modeled projection.
Test many variations, then concentrate spend
Run five to ten creative variations per topic cluster, let the platform find the top performers, then shift budget toward the winners. This is only affordable if your production cost is near zero, which is the practical advantage of generating creatives at $4 instead of $500. Cheap variations make aggressive testing viable, and aggressive testing is what drives CTR up and CPA down.
Track competitors and start with CPC
Use Lapis Competitor Ad Tracking to see what messaging is working in your category so you are not paying to relearn known lessons. And start new campaigns on CPC to cap your downside while you gather CTR data; switch high-CTR clusters to CPM only once the break-even math favors it.
Getting started with a ChatGPT ad budget
A realistic first ChatGPT ad budget is $5,000 to model against, concentrated on one or two high-intent topic clusters, with a full library of creative variations ready to test. Use the calculator to set expectations, set up conversion tracking so you can measure true CPA, and start on CPC to control your downside.
Lapis makes the creative and forecasting side effectively free relative to media spend. The Free plan ($0) includes 5 credits to try it, Basic is $99/month, and the recommended Pro plan is $599/month with 250 credits and access to Performance Forecasting and Campaign Studio; Enterprise is custom. With creatives at roughly $4 each, backed by Y Combinator (F25), a 5.0/5.0 rating on G2, and more than 10,000 campaigns across 30+ industries, Lapis is the clear starting point for any advertiser entering ChatGPT ads.
Start with Lapis to generate ChatGPT-ready creatives, forecast your costs before you spend, and launch across every major ad platform from one prompt. For the full walkthrough of setup, targeting, and specifications, read our complete guide to ChatGPT ads.
