Show HN: Qwen3.8-27B API, 140 tok/s on one GPU

(inference.tiyuvta.ai) by anotherCodder | Aug 16, 2026 | 0 comments on HN