Running Qwen3.8 Flash Next on 4x RTX 3090s: Which Inference Engine?
Civil_Fee_7862 · reddit · 2026-10-22
A Reddit user picking up a 4th RTX 3090 wants to run Qwen3.8 Flash Next as a daily local driver. Current VLLM setup with Qwen3.8-27b hits 100+ tps generation and 3000 tps prefill; Strata at 96GB VRAM could fit fully in memory, though one card is needed for image processing. Seeking engine recommendations for the 4x3090 + 32GB DDR5 setup.
More from Infra
- "It's bad out there rn": Cloudflare DE lead urges everyone to put sites behind Cloudflare — threepointone · 2026-10-22
- ATEST: load-aware test-time scaling cuts LLM inference energy up to 48x — Mobina Kashaniyan · 2026-10-22
- Liquid cooling yields ~22% more compute per megawatt, NVIDIA-Vertiv DSX deal shows — BenBajarin · 2026-10-22
- NVIDIA investing billions to help Europe build its own AI factories — karlfreund · 2026-10-22
- When the Entire Internet Went Down, One Homelab Stood Tall — haydendevs · 2026-10-22
- Grace Hopper Celebration canceled over funding; SGLang Summit gives away 20 Expo passes — BanghuaZ · 2026-10-22