OpenAI quietly ditched GPT-4.0 mini and we have questions
Matt V
May 16, 2025
OPENAI
OpenAI has just launched GPT-4.1 and GPT-4.1 mini inside ChatGPT.
The full version is now available to Plus, Pro, and Team users, while the mini version is open to both free and paid accounts.
As part of the update, GPT-4.0 mini is being phased out.
OpenAI says GPT-4.1 is better at following instructions and writing code than GPT-4o, and it’s also faster, making it a handy tool for developers.
While these models technically launched in April via API, this is the first time they’re being made directly available inside ChatGPT.
The launch didn’t come without some backlash.
When GPT-4.1 first dropped, some researchers criticised OpenAI for not releasing a safety report alongside it.
OpenAI pushed back, saying GPT-4.1 doesn’t count as a “frontier” model, it’s not more intelligent than GPT-4o and doesn’t add new features - so it didn’t need the same level of reporting. You be the judge.
In brief:
GPT-4.1 is built to be faster and better at code than GPT-4o.
It’s now in ChatGPT, and the mini version is free for all.
OpenAI is leaning into more transparency with a new safety hub.
Mini, but mighty
To address ongoing concerns, OpenAI has now launched a Safety Evaluations Hub, where it plans to share internal testing results more regularly.
The timing matters too, AI coding tools are having a moment.
OpenAI is reportedly close to acquiring Windsurf, a top coding assistant, for $3 billion.
And Google? It’s just upgraded Gemini to better integrate with GitHub.
GPT-4.0 mini, we hardly knew ye.