The article Chatgpt “Leave longer”: The Reasoning function in Fabian Peters test first appeared on Basic Thinking. You always stay up to date with our newsletter.
Chatgpt has received a new function with “Reasoning”. You can use a button to ask the AI to think longer before the answers. We took a closer look at the function.
After Deepseek, Chatgpt now also offers a reasoning function. Hot specifically: You can ask the voice model via a button to think about the answers longer. That costs a little more time, but should lead to better results.
Chatgpt: What is Reasoning?
The function goes back to the so-called Reasoning models from Openai, which are already in September 2024 were presented. The term stands for as much as “conclusion” or “logical thinking”. The Chatgpt’s Reasoning function is intended to simulate a cognitive thinking process in which various information processes, analyzed and evaluated.
As a user, you can partially understand this process. Because Chatgpt has a kind of soliloquy after you have promptly entered one. The voice model shows what it is doing (“superior”, “check”, “explain”). The process can take a few to several seconds.
The function is currently limited to the insertion and output of texts. Uploading pictures, videos or documents is not yet possible. The same applies to generating visual content. You can use the Reasoning function by in the Chatgpt input field on the Button with the light bulb Click. As soon as it shines blue, the AI thinks longer before the answers.
Reasoning of Chatgpt in the test
We took a closer look at the function to find out whether Chatgpt actually provides better answers in Reasoning mode. To do this, we have provided three different inquiries to the free standard model GPT-4-once with active starting and once without. Our three questions/entries:
- “My everyday work includes a lot of different tasks. How can I structure it to work more effectively? “
- “Summarize the following text so that a six -year -old child understands him.”
- “Why was the 2025 Bundestag election preferred in Germany?”
Conclusion: similar to worse answers
When asked about an effective workflow, Chatgpt spit out a table with seven different strategies in normal mode. Among them: the setting of priorities, the furnishing of themed blocks and the so-called 2-minute rule. It is about doing smaller tasks with a period of less than two minutes immediately.
In the Reasoning mode, Chatgpt thought five seconds longer. However, the result is similar. Once again, we received a table with approaches to help in everyday work. Instead of seven, however, only five strategies were described in more detail.
Since no documents can yet be uploaded in the Reasoning mode, we have manually inserted an article from us manually into the text field for the task for the text of the text. The result: the summaries were similar and both were relatively understandable. In the Reasoning mode, Chatgpt took ten seconds longer, but the reader-the six-year-old child-speaks directly to “you”/”Dir” in the answer. This can be an advantage for understanding.
When asked why the Bundestag election in 2025 was preferred in Germany, the answers were again very similar in terms of content. Striking: In normal mode, Chatgpt spat out and precise data on the day. This was not the case with Reasoning. The answer is a little worse.
Also interesting:
- “Fall in cognitive skills”: Artificial intelligence can make us more stupid
- AI Act: Anyone who uses AI in the company now needs certain skills
- Deepseek: China-Ki apparently consumes more energy than assumed
- Electricity and water consumption: the effects of AI on the environment
The article Chatgpt “Leave longer”: The Reasoning function in Fabian Peters test first appeared on Basic Thinking. Follow us too Google News and Flipboard.
As a Tech Industry expert, I believe that allowing the chat to “think longer” through the Reasoning function in the test can be an effective way to improve the overall performance and accuracy of the chatbot. By giving the chatbot more time to process and analyze information, it can generate more thoughtful and well-reasoned responses to user queries.
This feature can also help the chatbot make more informed decisions and provide more detailed explanations to users. In the fast-paced world of technology, it is crucial for chatbots to be able to think critically and logically in order to effectively assist users and solve complex problems.
Overall, incorporating the Reasoning function in the test can enhance the chatbot’s capabilities and make it a more valuable asset in the Tech Industry. It can lead to a more seamless and efficient user experience, ultimately improving customer satisfaction and trust in the chatbot’s abilities.
Credits