I doubt it, a lot of what I use llms for is on the very edge of what these models are trained for.