By skyforbes I've been testing version 5.1 since this morning and honestly, I think it still needs a lot of polish.
At first I found it an interesting improvement: the responses were longer, somewhat closer, it even reminded me a little of the tone of 4.0 in its heyday. But as the hours have passed, the experience has become exhausting.
At first I found it an interesting improvement: the responses were longer, somewhat closer, it even reminded me a little of the tone of 4.0 in its heyday. But as the hours have passed, the experience has become exhausting.
Each response seems written by a notary with excess enthusiasm: everything listed, in bold, with points and subpoints… and without any real need for so much embellishment.
If you tell him something everyday or pass on news, he immediately transforms: the warmth disappears and a cold, almost mechanical speech appears, more similar to Gemini than to the ChatGPT we knew.
I miss the naturalness, the balance between empathy and precision.
I don't need a thesis in each answer, but rather a fluid, human conversation, with rhythm and soul.
I hope they adjust the model soon, because in this state the 5.1 seems more concerned with sounding correct than connecting.
For now, I'll stick with the “normal” GPT-5, which is still the most human of all.