# 推理购买从"锁定供应商"转向"按需选择级别"

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-05-29 04:19
- AIHOT 分数：62
- AIHOT 链接：https://aihot.virxact.com/items/cmppyoycu00asslno4acy1at1
- 原文链接：https://x.com/rohanpaul_ai/status/2060093585727619292

## AI 摘要

The Grid AI 提出了一种新的AI推理购买模式。用户不再指定具体模型，而是根据任务复杂度选择标准（standard）、生产（prime）或极致（max）三个级别之一。平台会自动将请求路由到满足该级别要求的最便宜供应商。应用仅需接入单一API，后端模型可根据价格与质量动态变化，从而优化成本。作者曾用Hermes Agent在本地测试，通过agent-price级别处理了工单分类工作流。The Grid目前处于Beta阶段，声称通过供应商竞价可使AI API成本降低最高80%，并为新用户提供首200M tokens免费额度。

## 正文

Most AI teams still buy inference like they are buying software from 1 vendor.

They pick a model， accept the fixed price， wire it into the app， and keep paying that rate even when cheaper models could handle the same work.

@The_GridAI takes a different approach.

Instead of choosing a model name， you choose the level of work you need： standard， prime， or max. A simple task like support-ticket classification can run on standard. Normal production work like RAG， drafting， support replies， or agent steps can run on prime. Harder work with long context or higher error cost can run on max.

The Grid then routes the request to the cheapest supplier that still qualifies for that tier.

So the app still uses one API and mostly the same code， but the model behind the request can change as price and quality change.

I tested it with Hermes Agent on my Ubuntu machine. Hermes ran locally， while The Grid handled the inference through agent-prime. The workflow was simple： read support tickets， apply a policy file， and write a triage report.

### 引用推文

> The Grid：The Grid's Beta is LIVE! We can get your AI API costs down by up to 80% by making suppliers compete for your requests. Your first 200M tokens are on us, start b...