Post

I'm still working through the list of models to run through my own benchmark, but the preliminary results are in 😎 The new 2.6B model by Liquid AI works really well! Better than the 8B-A1B model, too!