GPT-4O, O3, O4-Mini … For some years, Operai has been offering several artificial intelligence models for Chatgpt users. The latter differ in relation to their abilities, especially in terms of reasoning. However, it was difficult for users to know which one to choose during their interactions with the chatbot.
The company itself realized, announcing last February its objective “to unify the models of the O series (reasoning models, editor’s note) and the models of the GPT series (flagship conversation models, editor’s note)” to simplify its offer. This is what Operai did on the occasion of the launch of GPT-5 on Thursday, August 7.
“GPT-5 is a unified system of an intelligent and effective model that answers most questions, a deeper reasoning model (GPT thought) for more complex problems,” explains the company in a blog article.
Simplification of the offer
Operai took advantage of the launch of GPT-5 to remove several of its old models (GPT-4O, GPT-4.1, O4-mini-High, O3-Pro …). “If you open a conversation using one of its models, ChatGPT will automatically change to the nearest GPT-5 equivalent,” the company said in a question frequently.
Therefore, conversations with models 4o, 4.1, 4.5, 4.1-mini, O4-mini or O4-mini-High will open automatically in GPT-5. Those with the O3 reasoning model will change to GPT-5 thought. If GPT-5 is intended for conversations, the latter allows “deeper answers” for reasoning.
Finally, the O3-PRO model is replaced by GPT-5 Pro. Only accessible to the subscribers of Chatgpt Pro (229 euros per month), this GPT-5 variant is intended for “more complex and demanding tasks”, thinking even more than GPT-5 thinking of providing “complete and higher responses”.
A choice made by a new tool
To the GPT-5 and GPT-5 thought models, a “mini” version for each one is added. Less powerful than the last one, but more than its removed equivalents, they will be used when users have reached the use limit.
This only refers to free users and subscribers of Chatgpt Plus (23 euros per month). As indicated by OpenAI, the first can send up to 10 messages to GPT-5 every five hours. Once the limit is reached, they will lean automatically in GPT-5 Mini. Similarly, they will only have a message per day for GPT-5 thought.
Chatgpt Plus subscribers can send up to 80 messages every three hours in GPT-5 and 200 messages to GPT-5 thinking per week before changing to mini versions. Having no limits in terms of use of the two models, the subscribers of Chatgpt Pro (229 euros per month) are not affected.
But in most cases, users, free and paid, will not need to think about which Chosir model. Because GPT-5 will take care of it instead.
By explicit intention, the company means that if a user says, for example, “reflecting on that” to their application, the tool will choose GPT-5 thought. Keep in mind that payment subscribers will also have the possibility to manually change from GPT-5 to GPT-5 thought and vice versa. Openai intends to simplify the use of its chatbot.
Source: BFM TV
