I've been watching ChatGPT's ad platform open up all year, and for most small businesses my read right now is simple: not yet. Six months ago the question was moot anyway. Ads were invite-only, with spending requirements that put the channel out of reach for smaller advertisers. That's no longer true. You can walk up to a self-serve platform and buy an ad today, provided you're in the right country and category.

With access no longer the main hurdle, I'm looking at the factors that still matter: cost, competition and whether ChatGPT ads are worth testing at all.

Why You Were Locked Out, and Why You Aren't Anymore

OpenAI announced its first ad test on January 16, 2026, for the Free and Go tiers, and ads went live February 9 for a hand-picked group including Adobe and Omnicom Media, per MediaPost. Small businesses were largely shut out by design: minimum commitments reportedly ran $200,000 to $250,000 during the early rollout, per Business Insider.

The door swung fully open May 5, when OpenAI launched a self-serve Ads Manager aimed at companies of all sizes. The minimum spend disappeared and CPC bidding arrived at $3 to $5 per click. AdRoll added another entry point, letting small business customers buy placements through their account manager.

What It Would Actually Cost You

OpenAI's pilot launched at a $60 CPM. Digiday reported that fell to around $25 by mid-April before minimums disappeared and OpenAI introduced CPC bidding at $3 to $5 per click.

That's a premium against what you're likely running elsewhere. Meta's average CPM is about $7.50, per WebFX, so ChatGPT impressions can cost considerably more. Per click, $3 to $5 lands closer to Google Search.

The bigger issue, and the thing that gives me pause, is what happens after you buy that traffic. Click-through rates on ChatGPT currently run well below Google Search, suggesting advertisers may have less opportunity to drive traffic at scale.

And a competitive cost per click doesn't tell you whether those clicks will turn into customers. Cost per outcome is the number that matters, and it's still unproven.

Who You'd Be Competing Against, and Whether It Works

The early roster was dominated by big names: Target, Adobe, Williams-Sonoma, Ford and Mazda, plus retailers brought in through Omnicom and Dentsu, per Digital Commerce 360. During the early self-serve rollout, more than 1,400 brands had tested ChatGPT ads by late May, per The AI Ad Economy. That still represented only about 1 percent of tracked brands, with many early advertisers operating on budgets far larger than a typical small business.

Early results cut both ways. Target saw ChatGPT traffic grow roughly 40 percent month over month, per Lengow, while Criteo found traffic referred from LLM platforms like ChatGPT converted at roughly 1.5 times the rate of other referral channels when it first disclosed the data, climbing toward 2 times in some retail categories by May. But WebFX put ChatGPT's overall click-through rate at 0.68 percent, far below Google Search's 4 to 6 percent. Nobody, including OpenAI, is claiming a proven benchmark yet.

Should You Actually Start Advertising on ChatGPT?

For most small businesses, that lands me back where I started: not yet, but watch it closely. Access is wide open, cost per impression is still higher than Meta, and nobody has shown this channel converts as reliably as what you're already running.

Expect brands to pile in over the next few quarters simply because access is easy now, not because performance is proven. That's normal for a new channel and doesn't mean you need to move first.

If your buyers already spend real time in ChatGPT, a small structured test, one campaign, one clear goal, is a reasonable way to learn the mechanics early. If your current channels are working and budget is tight, there's no penalty for waiting. The platform, pricing and measurement tools are all still changing month to month.