When you buy via links on our articles, Future and its syndication partners can win a commission.
Credit: Ilgmyzin / Unsplash
-
Openai has announced a new update for Chatgpt-4o
-
The update has improved the capacities of the model, which makes it “more intuitive, creative and collaborative”.
-
This completes a crazy week for Chatgpt after the launch of the native 4O images generation
Another day, another Cat update. This time, Openai has improved its GPT-4O model, making it “more intuitive, creative and collaborative”.
What exactly means? Updates publication notes highlight three major changes, which may be important or not depending on how you use Chatgpt.
The first improvement is in problem solving, where 4O is now smarter with regard to rod and coding problems. OPENAI says that GPT-4O “ generates the cleaner and simpler frontal code, thinks more precisely via the existing code to identify the necessary modifications and systematically produces coding outputs which compile and successfully run, rationalizing your coding workflows. ”
The second improvement is in the capacity of GPT-4O to follow the instructions as well as the precision linked to the formatting. With this change, Chatgpt should be better to follow your prompts with several requests – remember to make X, Y and Z, then to finish C.
The third and final improvement is probably the most interesting, because Openai says that this update will better understand the implicit intention behind their prompts, in particular with regard to creative and collaborative tasks.
Openai says you will see less emojis because 4o is now more concise and clearer, creating answers that are “easier to read, less congested and more targeted”.
Chatgpt “blurred” improvements
In AI, “blurred logic” is the ability of a model to understand and manage ambiguity and implicit intention behind an invitation.
I tested the GPT-4O update with a few simple prompts, and even if I had not yet seen a clear improvement compared to the previous model, the reference results do not lie.
According to the initial tests, it seems that this update has indeed improved the capacity of Chatgpt to manage the ambiguity, with a user on X sharing a Benchmark of creative writing Who sees GPT-4O exceeding Deepseek R1.
In lmarena.aiAnother reference used to determine the capacities of a GPT-4O model now exceeds the GPT-4.5 seated in Place number 2 behind Google newly launched from Google Gemini 2.5 Pro.
I think most people will not notice a difference here, but it is good to see that Chatgpt continues to improve, and with Openai adding the generation of native images to 4O Earlier this week is just the icing on the top.
What a crazy week for AI, next week cannot top it, right? RIGHT?