OpenAI launches new model that simplifies the ChatGPT experience

OpenAI announced that it has replaced the default model of ChatGPT with GPT-5.5 Instant. According to the information provided by the company, the new model produces more accurate results, especially in areas where the margin of error is expected to be low, such as medicine, law and finance. In addition, it is stated that the model’s tendency to produce false or fabricated information, called “hallucinations”, is reduced. GPT-5.5 Instant is available to all ChatGPT users as of today and replaces the existing GPT-5.3 Instant model. Within the scope of the update, not only the accuracy rate but also the way the answers are presented changes.

According to OpenAI’s statement, GPT-5.5 Instant shows improved performance in areas such as visual analysis, STEM-focused questions and determining when web search is necessary. While it is stated that the model gives more consistent answers, especially in technical content, it seems that shorter and more direct answers are targeted with unnecessary details reduced. Despite this, the company also emphasizes that ChatGPT’s characteristic conversation style has not been completely eliminated. In addition, reducing behaviors such as unnecessary emoji use, excessive formatting and constantly asking additional questions are among the notable changes.

With the new model, some innovations are also offered on the personalization side. It is stated that GPT-5.5 Instant is more successful in extracting context from previous chats, uploaded files and Gmail contents. Thus, more relevant answers can be generated based on users’ past interactions. However, these features are currently only available to paid Plus and Pro subscribers via the web version. While it is stated that support for mobile applications will come soon, it is stated that access will be expanded for Free, Go, Business and Enterprise users in the coming weeks.

Memory resources used in ChatGPT replies become more visible

Another innovation announced by OpenAI was the system called “memory sources”. With this feature, the user will be shown what past chats, files, or other contextual data ChatGPT is using when creating a response. Thus, it will be possible to see more transparently what information the model produces an answer based on. It can be said that this change is remarkable, especially in terms of understanding how the personal context formed during long-term use is evaluated.

In addition, OpenAI reveals that GPT-5.5 Instant is not only an update focused on speed and accuracy, but also aims to simplify the user experience. Recently, some users have been criticizing ChatGPT for giving too long answers, using emojis extensively, or asking unnecessary follow-up questions. It appears that the new model has adopted a more balanced communication style in line with this feedback. The preference for shorter and more focused answers, especially in professional usage scenarios, highlights this change.

On the other hand, it is not yet clear when the ChatGPT integration offered within Apple Intelligence will switch to the GPT-5.5 Instant model. No schedule has been shared by Apple regarding the update of the model currently used. Despite this, the fact that OpenAI has made the latest model update widely available indicates that the new version may spread to different platforms in a short time.