The research and the exact config I used to run the 125B Qwen3.8-Flash-Next on one RTX 3090 Ti with Strata, as an OpenCode coding model. Which quant and why, the install fixes Pop!_OS 22.04 needed, the sampling block Strata does not write for you, calibration results, a one-command switch back to llama.cpp, and a side-by-side speed probe against my old Qwen3.8-27B.