GPU advice for Qwen 3.5 27B / Gemma 4 31B (dense) — aiming for 64K ctx, 30+ t/s
Hey all, Looking for some real-world advice on GPU choices for running the new dense models — mainly Qwen 3.5 27B and Gemma 4 31B. What I’m targeting Context: 64K+ (ideally higher later) Speed: 30+ tok/s @ tg128 minimum Power: not critical, but lower …