**I pushed Qwen3.8-27B to 99 tps on a single request and 1150 on batch on a 3090**

> Qwen3.8-27B on a single RTX 3090: 416 tok/s batched or 25 ms/token single-user, 150k context, vLLM - syv-ai/qwen38-27b-rtx3090 Qwen3.8-27B on one RTX 3090 Serving setup for Qwen3.8-27B on a single 24 GB consumer GPU with vLLM. 150k token context, OpenAI-compatible API with key auth, and two ready-made configs depending on what you're doing: batch/ single-user/ for API backends, pipelines, many concurrent requests one...

**来源信息**
- **来源**:Hacker News:AI 热帖
- **分类**:论文
- **发布时间**:2026-08-18 04:05(北京时间)
- **原文**:[打开原文](https://github.com/syv-ai/qwen38-27b-rtx3090)