# 开放权重安全路径：Inkling 分级开放评估

- 来源：Nathan Lambert (@natolambert)
- 发布时间：2026-08-01 22:19
- AIHOT 分数：35
- AIHOT 链接：https://aihot.virxact.com/items/cmsagv4qi053jrob3ydcgq6yz
- 原文链接：https://x.com/natolambert/status/2083558154219229398

## AI 摘要

Thinking Machines 发布新博文，阐述开放权重模型的安全发布路径，主张在“无差别开源”与“少数实验室垄断”之间采取分级开放策略。文章以 Inkling 模型为例，说明其安全评估方法及分阶段扩大访问权限的理由。Nathan Lambert 认为此类工作对向公众解释开放模型持续发布的价值重大，有助于规避监管风险。

## 正文

This type of work is incredibly valuable for explaining to the broader public how you can keep releasing open models despite them having open ended risks.

Will protect against a lot of potential regulatory attention.

### 引用推文

> Thinking Machines：Releasing weights indiscriminately isn't safe. Neither is keeping capable models inside a few labs. We think there's a path between them. We haven't mapped all ...