ChatGPT 4 is vastly superior to ChatGPT 3.5 in my experience, although it still fails at logic sometimes.
Mixtral 8x7B. Short of that (high ram requirements), I've also found Mistral v0.2 to be pretty solid for a 7b model.
YMMV though - it's going to depend on your use cases.
Do you have instructions and RAM requirements for running that model? Llama.cpp?
I just use Ollama[1] - makes it incredibly easy to get going on MacOS, and can be run on linux/WSL also. RAM required will depend, but generally to run Mixtral at reasonable quantization levels (e.g. Q4) you're going to want 36GB or more.