RTX 3090 Runs Qwen 27B Locally at 100 Tokens/s

Users demonstrate running Qwen3.8-27B (INT4) locally on a single 24GB RTX 3090 with vLLM, achieving 144K context and up to 100 tokens/s—a big speedup over MacBook setups.

2026-09-14 ~ 2026-09-14 · 2 related posts

• Qwen3.8 27B INT4 With 144K Context Runs on a Single RTX 3090 via vLLM — Altruistic_Heat_9531 · 2026-09-14
• RTX 3090 local setup runs Qwen 27B at 100 t/s, 5x faster than MacBook M5 Pro — gutard · 2026-09-14