3 million GPTs later, OpenAI's storefront finally opens

Two months after letting users build custom AI agents, OpenAI is finally opening the doors to its GPT Store — and the sheer volume is staggering. More than 3 million custom GPTs have been created since the feature launched, and now paying users on ChatGPT Plus, Team, and Enterprise plans can browse a curated storefront to find ones that are actually useful. That’s a lot of noise to cut through, and the company is trying to surface signal with a community leaderboard, category filters spanning DALL·E, writing, research, programming, and education, and a handful of featured picks from partners like Khan Academy, Canva, and AllTrails.

If you’re a builder hoping to get your GPT listed, there’s a new hurdle: verification. You’ll need to save your GPT for “Everyone” and verify your builder profile with either a name or a website. OpenAI is also layering on a new review system that combines human and automated oversight, with a reporting mechanism for users to flag problematic bots. That’s on top of existing safety measures baked into the product. It’s a necessary step — an unmoderated firehose of 3 million GPTs would be a liability, not a feature.

But the real hook for developers is coming later. In Q1, OpenAI plans to launch a revenue-sharing program where U.S. builders get paid based on user engagement with their GPTs. They’re light on specifics right now — no word yet on the exact criteria or payout structure — but dangling a paycheck could transform this from a fun side project into a legitimate platform play. It’s the same playbook that turned app stores into gold rushes.

The store also splits into private sections for Team and Enterprise customers. Teams get a workspace-specific area for securely published internal GPTs, and Enterprise customers are getting beefed-up admin controls to govern which external GPTs employees can touch. Crucially, OpenAI reiterates that it won’t train models on conversations from those paid plans. For enterprises still skittish about data leakage, that’s table stakes — and they’re saying it out loud.