Hacker News · AI·4d agoRun Qwen 3.8 Flash Next (125B) on consumer hardware (RTX 4090) at 100T/s#qwen#inference#rtx-4090AI tools & infra1