The models AI researchers develop are cutting edge, and mathematically involved; the kind of thing I would expect an LLM to fare poorly at. And today's researchers are capable of implementing their models in pytorch and the like.