Spring Cleaning in February: OpenAI Clears the Model "Zoo" Out of ChatGPT

It seems OpenAI has decided that democracy and wide choice are all well and good, but order matters more. If you are used to juggling different versions of neural networks in search of the perfect answer, I have news for you. The company has declared February 13 a day of "general cleaning." And judging by the list of candidates for removal, Sam Altman's new broom is a harsh one.

What is happening?

It is like a restaurant menu that has grown so large that guests spend more time choosing a dish than eating it. OpenAI is removing a whole layer of models from the ChatGPT user interface. On the chopping block: GPT-5 (in its Instant and Thinking variants), plus the veterans of the stage: GPT-4o, GPT-4.1, GPT-4.1 mini and o4-mini.

Let's be honest: did you really understand the fine difference between GPT-4.1 mini and o4-mini in everyday tasks? Most likely not. Most users simply clicked the "smartest" or "fastest" button. Now that choice will be made for you.

Why now?

There are three reasons, and none of them is that the models are "tired."

First, inference cost. Keeping a dozen different sets of model weights hot and available for millions of users burns money on an industrial scale. Each model needs its own infrastructure, caching and request routing. Simplifying the lineup is classic P&L (profit and loss statement) optimization.

Second, the paradox of choice. UX research has said for years that if you give a user 10 options, they freeze. Give them one, but the best, and they are happy. OpenAI is clearly moving toward a "single window" concept, where the system decides how much compute to spend on your query without asking whether you want the Thinking or the Instant version.

Third, data focus. To train the next generations (GPT-6?), the company needs you to use current models. Your likes and dislikes on answers from outdated versions are worth far less for RLHF (reinforcement learning from human feedback) than feedback on the flagships.

What about developers?

Here OpenAI showed rare prudence. The changes affect only the "civilian" ChatGPT. Those who build their apps through the API can breathe out. Nothing in their code needs to change, at least for now. That makes sense: breaking production for enterprise customers the day before Valentine's Day would be too cruel even for a tech giant.

Usually such a gap between versions in the UI and the API signals that the old models are being moved to "legacy" status. They will remain available to machines but hidden from human eyes, so as not to spoil the impression of the product.

The bottom line:

The era of endless version choice in chatbots is ending. OpenAI is clearly hinting: stop micromanaging models, just ask questions. Are we waiting for the announcement of a single universal model that will replace this whole zoo?