Together AI 详解开源 AI 技术栈：模型、推理、路由、Harness 与工具
原标题：The Open Source AI Stack: models, inference, routers, harnesses & tools →
AI 摘要
Together AI 发布博客，系统梳理开源 AI 技术栈，将其拆解为模型、推理、网关与路由、Harness、工具（Skills 与 MCP）五个相互独立的层次。文章指出开源模型质量已接近闭源模型，开发者无需自行训练模型或购置 GPU 即可迁移，并强调分层解耦让开发者能快速切换新模型。文中以 Kimi K3（1.8T 总参数、104B 激活参数）和 GLM 5.3 Flash（320B 总参数、18B 激活参数）为例，说明大模型适合复杂模糊任务、小模型适合明确窄任务，后者约小 6 倍、便宜 20 倍。
正文节选
As the quality of open source models have bridged the gap with closed source models, a lot of developers and organizations are looking to move to open source models for more ownership, control and economics. This post is a deep dive into the open model AI stack that developers need to consider as they move from closed to open source. Using open models for agentic software development does not require learning how to train models, buying a rack full of GPUs, or becoming an expert in machine learn