GPT-6 Sol gets work done but is tiring to talk to, one user says
eidzoku says GPT-6 Sol feels more like 5.2/5.3 Codex than 5.6, struggles to read intent and casual chat, and is considering Claude Max to use Sol only as a subagent.
I'm trying to like GPT-6 Sol but I'm not sure if I'm gonna make it😭 In terms of of communication it feels more like 5.2\5.3 Codex rather than 5.6 Most likely the model is smaller, but dear lord it's insufferable when it comes to understanding intention, taste... damn, even casual chatting. I won't say 5.6 was perfect, but it was better than Opus 5, on the other hand this... 6 thing. It gets shit done. But it's just tiresome to work with. You can banter with Opus a lot, discussing your idea and implementation or you can banter a lot trying to explain something to GPT again and AGAIN. I hate the fact that the most toxic AI lab in US knows how to make pleasant to talk model and OpenAI don't. Well, maybe they do, but intentionally making their models as 'autistic' as possible. I feel an urge to buy Claude Max and use Sol as a subagent idk........