GPT-5 solved an important problem for me: which model to use in daily life. At this stage of model development, the performance improvement is no longer exponential, and in most cases, it is not bad to use. So with GPT-5, my usage may become: -Claude Code with Opus 4.1 is responsible for all coding scenarios -Leave all other scenes to GPT-5