▲ 1 pointsShow HN: Qwen3.8-27B API, 140 tok/s on one GPUinference.tiyuvta.aiby anotherCodder·10h ago·0 comments·view on hn ↗