I just copied and pasted it, can't do markdown.

https://pastebin.com/FAdBbGUd Qwen3.8-27B on an RTX 3090: Inference Options and Benchmark Results