Openai has changed the way its new chatgpt O3-mini model shows its tank chain (COT) to “make it easier for people to understand how the model thinks,” but it has been exposed to an almost immediate setback from users and accusations that it copies the way in which Deepseeks R1 model shows its reasoning.
In a speech with Techradar, Openai said, “Users have told us that understanding how the model reasons through an answer not only support more informed decision making, but also helps to build confidence in its answer.”
“While the model’s raw cot remains hidden as it is difficult to understand, we have found a balance: the model can think freely, and then it organizes these thoughts so that they are easy to read. To improve clarity and security we have added An additional finishing step where the model reviews the raw thought chain, removes any uncertain content and then simplifies all complex ideas.
The new approach effectively provides summary of the model’s reasoning instead of showing you the raw data. Despite the new approach available in both O3-mini free and paid users and O3-mini-high for paid users, the answer to X on its changes was not quite positive.
Mayo Oshin replied to X and posted, “We would appreciate if you showed the full thought chain, not just summarized version …. thank you”, and Conor replied with, “It’s still a summary and not really cot, like is disappointing. “
Some other users responded by suggesting that Openai simply responded to the threat offered by the new Deepseek by copying the way it presented the reasoning chain in its R1 model. “Finally, deeply changed the o-world for us,” Hamza replied. “So Openai copied Deepseeks chain of thinking function?” said Ignis Rex, and “the moment China is what is innovating and the United States is copycat,” Josip Tomo Licardo said.
A big step forward
Personally, I very much prefer the new way of presenting reasoning in chatgpt O3-mini. If you try the model now (just press the reasoning button) before entering your prompt, you will find that there is much more information about how Chatgpt comes to its conclusion. The previous lack of reasoning information compared to Deepseek was something I criticized 03-mini in the last few days. So from my perspective, I think the new approach is a big step forward for chatgpt O3-mini.
I am also not alone in welcoming the new feature. X uses novel Pshichenko said, “Hilary, the thought chain is more adorable and funny than the stiff response,” and “It’s a nice improvement! I would very much like to see some examples of how this update affects its justification process, “said Jesus Vazquezon X.
