> They are refining it to perform better at certain tasks.

Like what, coding?

Over 50+ LLMs out there do exactly this.