The biggest decline after Openai announced that its new GPT-5 model for Chatgpt came from faithful older models who thought that the new generative generator chatbot did not have the plume of its predecessors.
Now you have more choice of pre-GPT-5 models (although you will have to drive some of them) and better control over GPT-5 components manages your questions.
Openai is still in the process of sorting a somewhat rocky launch of GPT-5, led by complaints concerning the lack of choice of models. The model has been anticipated for more than two years and comes from competitors like Anthropic and Google have published new powerful versions of their AI models this year. (Disclosure: Ziff Davis, CNET’s parent company in April, filed a complaint against Openai, alleging that it has violated Ziff Davis Copyrights in the training and exploitation of its AI systems.)
OPENAI plans a model that could manage everything: GPT-5 includes two different modes, a quick and lean for simple tasks, and a viewable to reason for complicated tasks. A routing program would decide which model has managed a given prompt. It is always the default chatgpt value, but it is not your only option.
Look at this: Chatgpt users want old models to come back, Intel PDG goes to the White House and more | Tech today
Choice of GPT-5 models
There are a few different modes of GPT-5 that you can select between if you want to use the new OpenAi technology. Here is a quick overview:
- Auto: This mode allows the switching program integrated into GPT-5 to decide whether your request is managed by a lighter and faster tongue model or a larger and slower model of reasoning. OPENAI CEO Sam Altman Posted on X that it will be best suited to most people.
- Fast: Your request will go directly to the fastest and lightest model. Expect quick and basic responses, but not in-depth research that you will get from a reasoning model.
- Thought: This is a reasoning model, which means that it will try to answer your question on several stages. He can use web research and other tools, or he could go back and redo his past steps to try to get the right answer. There are a few limits on the amount you can use this model (3,000 messages per week at the moment).
- Think mini: It is a lighter and faster version of the reflection model. If you reach the line use limit, you will be stuck with Thinking Mini. You can access it for yourself if you activate more model options (more about this below).
- Pro: The most powerful reasoning model on the fleet is only available until the moment for professional users, who pay $ 200 per month. (Although Altman has teased giving more users, who pay $ 20 per month, a very limited number of requests as a taste.)
Widen the image
Here is what your choices will look like if you do not activate the models inherited in the settings.
Again, sticking to Auto is probably easier for most users. Think about it as driving a car with an automatic transmission instead of having to change speed manually. Of course, people who are really in cars might prefer the change of stick, but most people should probably let the machine manage it.
How to get older Openai models
Anyone with a paid chatgpt subscription can access the old GPT-4O model directly in the same menu where you can choose your GPT-5 flavor. This model received the most clamor from users after its deletion, and Altman said that if Openai never decided to remove it permanently, “we will give a lot of notice.”
But 4o is not your only choice (if you are a paying user). Just know where to look for. To access GPT-4.1, 4o-mini and 3o, as well as GPT-5 Thinking Mini, you will have to enter your settings and switch to “Display additional models”.
A model, many personalities
The Openai plan was for a GPT-5 model to replace all the ancients, and that always seems to be the goal. Altman said that changes to default personality De GPT-5 should satisfy certain dissatisfied users of the new version. GPT-5 also allows you to personalize the way it reacts to you. If you want it to be more sassy or more cheesy, you don’t need to change models for this. This seems to be the long -term goal.
“The real solution here remains to allow users to personalize the much more chatgpt style,” said Altman said on X.
