Comment by waNpyt-menrew

Larger model, better benchmarks. Bigger bomb more yield.

Any benchmarks where we constraint something like thinking time or power use?

Even if this were released no way to know if it’s the same quant.