Qwen3.8 27B is live on Together AI for fine-tuning and dedicated model inference — train on your data and serve on dedicated infrastructure without stitching stacks.