Ive been using gpt 5.4 and 5.5 and honestly 5.4 is solving everything at the pace I need it. I'm the biggest bottle neck in terms of reviewing PRs and my own co

Ive been using gpt 5.4 and 5.5 and honestly 5.4 is solving everything at the pace I need it. I'm the biggest bottle neck in terms of reviewing PRs and my own code. So having a model which can solve a complex task in 10 minutes vs 30 minutes doesn't really give me any meaningful improvement. Also, the biggest factor is having a good planning phase. A good plan is better than even major model improvements.