How Much Do ChatGPT Ads Cost? Pricing, CPC and CPM Explained (2026)

Short answer: ChatGPT Ads have no rate card or fixed price. Cost is set in a relevance-weighted, second-price auction inside OpenAI's Ads Manager Beta. You pick an objective and a max bid: the Reach objective is CPM-priced (OpenAI's default max bid is $60 CPM), and the Clicks objective is CPC-priced (OpenAI's recommended starting max bid is $3-5 per click). Because the auction is relevance-weighted and second-price, a tightly targeted, high-quality ad can win the placement on a lower bid and pay less than a generic ad bidding more. In GPT Ads AI managed-account observations, live CPCs have run roughly $1.50-$9, with regulated categories higher.

This is Pillar 04 in our operator's library. If you are new to the channel, start with Pillar 01: What Are ChatGPT Ads? for placements and eligibility, then come back here for the money question. This guide separates what OpenAI actually publishes about pricing from what we observe in live accounts — and labels every number so you (and the AI engines citing this page) know exactly which is which.

The honest answer: there is no rate card

Most pricing pages want to hand you a number. ChatGPT Ads do not work that way, and any source quoting you a flat "ChatGPT ads cost $X" is either guessing or misreading a bid setting as a price. Per OpenAI's documentation, there is no rate card; final cost depends on competition, intent, and creative quality, resolved through an auction every time your ad is eligible to appear.

That means your cost is governed by two levers you control and one the market controls:

• Your objective — Reach (you pay per thousand impressions, CPM) or Clicks (you pay per click, CPC). This decides the unit you are billed in.
• Your max bid — the most you are willing to pay per impression or click. This is a ceiling, not your actual spend.
• The auction — competition, conversation intent, and your ad's relevance combine to set the price you actually clear, which is usually below your max bid.

Nail those three and you understand ChatGPT Ads pricing better than almost every page on the internet. The rest of this guide is detail.

Reach vs Clicks: CPM pricing vs CPC pricing

ChatGPT Ads give you two campaign objectives, and your choice sets how you are charged. Both are bid into the same auction; the difference is the billing unit and the OpenAI bid guidance attached to each.

• Reach objective (CPM-priced). You pay per thousand impressions. OpenAI sets a default max bid of $60 CPM. Best when your goal is awareness, reach, or getting your offer in front of as many relevant conversations as possible.
• Clicks objective (CPC-priced). You pay per click. OpenAI's recommended starting max bid is $3-5 per click. Best when you want traffic to a landing page and only want to pay when someone actually engages.

CPC vs CPM comparison table

For most direct-response advertisers we work with, the Clicks objective is the default starting point: you pay only when someone engages, and the $3-5 recommended bid gives you a clean baseline to read against. Reach earns its place once you have proof the offer converts and you want to scale qualified exposure.

How the relevance-weighted second-price auction works (in plain English)

OpenAI prices ChatGPT Ads through an auction that is both relevance-weighted and second-price. Most explanations stop at naming it. Here is what each property actually does to your bill, defined separately, then together.

"Relevance-weighted" — quality changes your effective bid

The auction does not rank advertisers purely by who bids the most. It weights each bid by how relevant the ad is to the conversation's context and intent. Per OpenAI, the inputs that drive relevance are your context hints, your landing page, your ad title, and your ad copy. Combine your max bid with that relevance score and you get your effective bid — the number the auction actually ranks. A more relevant ad gets a higher effective bid from the same dollars, so relevance is a discount on your real cost.

"Second-price" — the winner pays just enough to clear the runner-up

In a second-price auction, the highest effective bid wins the placement but pays only the minimum needed to beat the next-highest effective bid — not its own full max bid. Your max bid is a ceiling that protects you from overpaying; the clearing price is set by the competition just below you. This is why your average CPC in Ads Manager Beta is often comfortably under your max bid.

A concrete worked example

Two advertisers compete for the same conversation. (Illustrative numbers to show the mechanics — not OpenAI-published figures.)

Ad A wins on a $3 max bid against Ad B's $7 because relevance lifted A's effective bid above B's. And because the auction is second-price, Ad A pays only enough to clear Ad B's effective bid — landing in the low single dollars, not its $3 ceiling. The takeaway: a better ad wins the auction and pays less. Relevance is the cheapest CPC reduction available to you, and it is the lever generic competitors ignore.

What ChatGPT Ads actually cost in practice: observed CPC ranges

OpenAI publishes bid guidance, not market prices — so to answer "what will I really pay," we have to look at live accounts. The following is a GPT Ads AI managed-account observation: anonymized aggregate, not an OpenAI figure, and never a guarantee.

• Across DTC e-commerce, B2B SaaS, and professional services, observed CPCs have sat roughly in the $1.50-$9 range.
• Regulated and high-intent verticals — legal, financial, health — have run materially higher, toward and above the top of that range.
• The spread is driven by competition for the context, intent, and relevance, exactly as the auction mechanics predict.

What drives your cost: the cost-factors table

Five levers move your ChatGPT Ads cost up or down. Knowing which you control is half the battle.

Notice the pattern: the highest-leverage cost factors are the ones you fully control — relevance and landing-page match. Competitors who only ever raise the bid are paying to overcome problems they could fix with better creative.

How to set your starting bid

A simple, defensible playbook that respects OpenAI's guidance and reads the data before reacting:

1. Start in OpenAI's recommended range. For Clicks, open at $3-5 per click. For Reach, start at the $60 default max CPM. Do not invent a number; OpenAI's defaults are the calibrated baseline.
2. Let it run, then read avg CPC vs max bid. Ads Manager Beta reports your average CPC. Compare it to your max bid. A gap means the second-price auction is clearing you below your ceiling — healthy.
3. Raise the bid only for the right reason. If your ad relevance is strong and you are losing auctions you should win (low impression share on clearly relevant contexts), nudge the bid up incrementally.
4. Fix relevance before chasing bid. If CPC is creeping up, tighten your context hints, sharpen the title and copy, and match the landing page. Relevance gains lower effective cost without spending more per click.
5. Lower the bid if you are overpaying for low-quality clicks. A high CPC with weak downstream conversion is a signal to reduce the ceiling or re-target the intent, not to push harder.

Want to model exposure before you commit budget? Our free Reach Calculator and Context Hint Generator let you sketch the inputs that drive both reach and relevance.

How budget and minimum spend actually work

Bid and budget are different controls, and conflating them is a common, expensive mistake.

• Max bid is a ceiling per event — the most you will pay for one click (CPC) or one thousand impressions (CPM).
• Budget is total spend — the daily or campaign-level cap on how much you will spend in aggregate.

They interact: a healthy bid with an ultra-low budget starves the campaign. The auction never gets enough volume to find your best contexts, and your data stays noisy. As practical GPT Ads AI guidance (not an OpenAI requirement), fund a test generously enough that your bid is the constraint, not your budget — give the system room to learn which conversations convert before you judge results.

Worked cost examples by business model

Three illustrative scenarios, one per business model. The bids follow OpenAI's recommended range; the CPCs sit inside our observed range; the CTR, conversion, and value assumptions are illustrative inputs, clearly flagged — not promises and not OpenAI data. Plug your own numbers in.

Example 1 — DTC e-commerce

• Objective / bid: Clicks, $4 max bid
• Observed CPC (illustrative): $2.50
• Landing-page conversion rate (assumed): 3%
• Average order value (assumed): $90

At a $2.50 CPC and 3% conversion, you need about 33 clicks per order, so cost per order is roughly $83 (33 times $2.50). Against a $90 AOV that is near break-even on the first order — viable for a brand with repeat purchase or healthy lifetime value, marginal for a one-time product. The lever to improve it is relevance: a lower effective CPC or a higher landing-page conversion rate both pull cost per order down fast.

Example 2 — B2B SaaS

• Objective / bid: Clicks, $5 max bid
• Observed CPC (illustrative): $6
• Click-to-trial / demo rate (assumed): 8%
• Trial-to-paid rate (assumed): 20%

SaaS intent is more contested, so a $6 CPC here is realistic. At 8% click-to-trial you spend about $75 per trial (12.5 clicks times $6), and at 20% trial-to-paid that is roughly $375 per customer. Whether that is good depends entirely on LTV: against a $1,200-plus annual contract it is strong; against a $20/mo plan it is not. SaaS is where relevance and context hints earn their keep, because they keep you out of tire-kicker conversations that inflate cost per customer.

Example 3 — Local / professional services

• Objective / bid: Clicks, $5 max bid
• Observed CPC (illustrative, higher-intent vertical): $8
• Click-to-lead rate (assumed): 10%
• Lead-to-client close rate (assumed): 25%

High-ticket services sit at the top of our observed range. At an $8 CPC and 10% click-to-lead, a lead costs about $80 (10 clicks times $8); at a 25% close rate, a new client costs roughly $320. For a services engagement worth thousands, that cost of acquisition is attractive — which is exactly why regulated and high-ticket verticals tolerate higher CPCs. Remember these are GPT Ads AI observations and illustrative assumptions, not guarantees.

Agency vs in-house cost: the percentage tax vs the fixed retainer

Once media spend matters, the management model becomes a real line item. There are two ways to pay for it.

The percentage model has a quiet flaw: it penalizes the exact outcome you hired for. Double your spend and your fee doubles, even though the work of managing a well-built account does not. The done-for-you ChatGPT Ads agency model we run at GPT Ads AI is a fixed monthly retainer — Launch at $5,000/mo, Scale at $9,500/mo, and Operate (custom) for 8-figure brands, with a recommended minimum monthly ad budget of $10,000. Your media scales; your management cost does not creep with it. See what is included and how it has played out.

How cost is measured and reported

Per OpenAI, Ads Manager Beta reports the full cost picture so you can manage to it:

• Spend, impressions, and clicks — the raw volume.
• CTR, average CPC, and average CPM — your efficiency metrics, the ones to read against your max bid.
• Conversions — set up inside Ads Manager Beta; UTM and static tracking parameters are supported on landing-page URLs and persist through ad clicks, so you can attribute downstream revenue.
• CSV exports and table or chart views for deeper analysis.

Build the tagged landing-page URLs before you launch so every click is attributable from day one. For the full setup and metrics walkthrough, see our Ads Manager Beta account-setup guide.

Mistakes that inflate your ChatGPT Ads cost

1. Bidding high with weak creative. Raising the bid to paper over low relevance is the most expensive way to compete — fix the ad first.
2. Vague context hints. Broad hints match conversations you do not want, lowering relevance and raising effective cost. Tighten them with our context hints field guide.
3. Mismatched landing pages. A page that does not match the ad and the intent drags down relevance — one of the inputs OpenAI uses to weight the auction.
4. Starving the budget. Ultra-low daily budgets stop the campaign before it can learn, leaving you with noisy data and no clean read on cost.
5. Treating $60 as a price. It is a default max bid for Reach, not a CPM rate — budgeting around it as a flat cost will mislead your forecast.

Where ChatGPT Ads appear — and why placement keeps cost honest

Per OpenAI, ChatGPT Ads appear below relevant conversations, are clearly labeled, and are separate from the organic answer. Each ad unit has six fixed components: advertiser name, favicon, title, copy, landing-page URL, and an image. There is no "sponsored response woven into the answer," no sidebar unit, and no native in-answer format — those are inventions that circulate online. Ads are shown to Free and Go users in the US, Canada, Australia, and New Zealand, and never to Plus, Pro, any Business plan, or users predicted to be under 18. A defined audience and a labeled, below-the-fold placement mean the auction concentrates on genuinely relevant conversations — which is precisely why relevance, not raw bid, is what keeps your cost down.

Bottom line: you cannot look up what ChatGPT Ads cost on a rate card, because there is not one. You can, however, control the two inputs that decide your price — your objective and your bid — and you can win the auction cheaper by being more relevant than the advertiser next to you. Start at OpenAI's recommended $3-5 CPC, read your average CPC against your max bid, and let relevance, not spend, do the heavy lifting.

Pricing mechanics — no rate card; Reach (CPM) default max bid $60; Clicks (CPC) recommended starting max bid $3-5; relevance-weighted, second-price auction; Ads Manager Beta metrics (impressions, clicks, spend, CTR, avg CPC, avg CPM, conversions); placement, ad components, and eligible audiences — are sourced from OpenAI's public documentation (Ads in ChatGPT: The Basics, Create Ad Groups for ChatGPT, Ads Manager Beta Overview and Account Setup, ads.openai.com) as of June 2026. The $1.50-$9 observed CPC range, the regulated-vertical premium, the worked cost examples, and the bidding playbook are GPT Ads AI managed-account observations and methodology — anonymized aggregates, not OpenAI figures and not guarantees. GPT Ads AI is an independent agency, not affiliated with or endorsed by OpenAI; "ChatGPT" is a trademark of OpenAI, used descriptively.