Old RTX 3090s Run Qwen3.8-Flash-Next at 1200 tok/s

QuixiAI demonstrated Qwen3.8-Flash-Next-NVFP4 running at up to 1200 tok/s with 262K context on eight five-year-old RTX 3090s, and about 120 tok/s per card on a quad-3090 setup via SlimServe.

2026-09-06 ~ 2026-09-07 · 2 related posts

• Serving Qwen3.8-Flash-Next at 262K context on 8 RTX 3090s hits 1,200 tok/s — QuixiAI · 2026-09-06
• Qwen3.8-Flash-Next-NVFP4 hits 120-361 tok/s on Quad 3090 via custom P2P driver — QuixiAI · 2026-09-07