Which model should I try?

In my current workflow (coding in python/c++ and technical reports) I mostly use Qwen3.6 27B and Gemma4 31B. In the past I tried other models like Deepseek with decent results but was painfully slow. so do you think there is some model that I'm missing and should try? EDIT: to be clear, I'm not asking how to make those models run faster, I'm asking which other models I should try.