User pushes Qwen 3.6 27B to 262K context with good results

A Reddit user reports successfully running Qwen 3.6 27B at up to 262K tokens of context on RTX 3090 Ti with Rope/Yarn scaling, maintaining coherence. They seek community input on higher context lengths.