Devs - do you use Mistral Medium 3.5 (128b dense) and if so - thoughts?

I've picked up a 3-bit quant of this one (Unsloth - Q3_KS) - the best fit for my config right now. Normally I shy away from 3-bit quants but as this is a giant dense model, I figured... why not. I've tested it a few hours in my latest project, and it found a few things that my daily driver had missed. It seems really good. Slower than an MoE, but not bad for me (8 tok/sec - with KV 80k - quant K = q8_0 quant v = q5_0). I was wondering what other people thought after actually using it with code. submitted by