Qwen3.6-35B becomes competitive with cloud models when paired with the right agent

A short follow-up to my previous post, where I showed that changing the scaffold around the same 9B Qwen model moved benchmark performance from 19.11% to 45.56%: https://www.reddit.com/r/LocalLLaMA/s/JMHuAGj1LV After feedback from people here, I tried little-coder with Qwen3.6 35B. It now lands in the public Polyglot top 10 with a success rate of 78.7%, making it actually competitive with the best models out there for this benchmark! At this point I’m increasingly convinced that part of the