OpenAI quietly dropped new API pricing and the developer corner of the internet basically detonated overnight.
The headline number everyone's screaming about?
GPT-4o mini is now sitting at 15 cents per million input tokens and 60 cents per million output tokens.
For anyone who's been paying GPT-4-tier rates, this is the moment your side project stops eating your lunch money.
The older, chunkier models still cost real money, but the mini tier is now cheap enough that you can burn tokens on dumb ideas and not hate yourself.
Here's why this actually matters beyond the sticker shock.
Cheaper tokens mean builders can finally ship the stuff they've been dreaming about but couldn't afford to run at scale.
Think AI study buddies, customer support bots, that app that roasts your outfit pics.
Cheap doesn't mean free, and rate limits still exist.
If your app goes viral and suddenly 50,000 people are hammering your endpoint, you're still on the hook for the bill.
Some devs are calling this a straight-up price war with the other big AI labs, and honestly the timing tracks.
Everyone's racing to lock in developers before the next shiny model drops.
Whoever gets you building on their API now probably keeps you for years.
So should you actually switch your stack?
If you're already deep in another ecosystem, migrating is a whole thing and maybe not worth the headache.
But if you're starting fresh or your margins are getting cooked, this is a genuinely strong reason to give it a look.
The bigger vibe here is that AI is getting commoditized in real time.
When the same capability costs a fraction of what it did a year ago, the moat stops being the model and starts being your product.
That's the actual story nobody's tweeting about.
This is great news for indie builders and terrible news for anyone whose entire pitch was "we use the expensive model." The floor just dropped out, and the people who win are the ones who build something people actually want.
Final Thoughts
Cheap tokens are a gift, but they're not a business.