Ugh I didn't want to post this... Back to Qwen3.8 27B

I don't know if anyone else has ran into these issues but when using Qwen Flash Next, my confidence in it at iq4_xs is high but not 100%. I notice it not following instructions, hallucinating more often and surprisingly it uses way less tokens than 27b. After using Strata.... yes i know.... I thought it was a breath of the next step in Ai. I was mistaken, yes it is good, yes it is fast. Yes it can do better than 27b in some circumstances... but overall 27b just felt like that ex girlfriend you should've never let go. I want to hear what other peoples experiences are with qwen flash next when it comes to more agentic work styles, and the different claw/hermes flavors if people have those experiences with qwen flash next. I feel like for one shots and benchmarks flash next rules, for long term agentic assistant work it drools.. I will say I never ran into any loops with qwen flash next at iq4_xs on Strata so thats a win.