Qwen 3.5 397B vs Qwen 3.6-Plus

I see a lot of people worried about the possibility of QWEN 3.6 397b not being released.

However, if I look at the small percentage of variation between 3.5 and 3.6 in many benchmarks, I think that simply quantizing 3.6 to "human" dimensions (Q2_K_XL is needed to run on an RTX 6000 96GB + 48GB) would reduce the entire advantage to a few point zeros.

I'm curious to see how the smaller models will perform towards Gemma 4, where competition has started.