Show HN: Qwen3.8-27B API, 140 tok/s on one GPUinference.tiyuvta.ai1 point·anotherCodder··0 commentsOpen articleSaveView on HN