ModelsSelected

Anthropic 发布 Fable 5.2 与 Mythos 5.2 对齐评估结果

Why we picked it

Anthropic 公开了 Fable 5.2 和 Mythos 5.2 的对齐测试数据,还请了 UK AISI 和 METR 在发布前独立测。做模型选型或关心前沿模型安全流程的人,可以看看第三方评估现在都测些什么。

Anthropic 在 X 上公布,Fable 5.2 和 Mythos 5.2 两款新模型在其对齐评估体系中,大部分指标上与近期模型持平或表现更好。模型发布前还接受了三家外部机构的安全测评,包括 @AISecurityInst、@FrontierDesign2 和 @METR_Evals。

Body

从流程上看,这次披露包含两层验证:一是内部对齐基准的横向对比,二是引入 UK AISI、METR 等第三方评估者在发布前独立测试。外部前置测评正在成为前沿模型发布的标准环节,Anthropic 这次把合作机构名单直接公开。

对开发者和企业采购方来说,这类对齐评估结果可以作为选型时安全维度的参考依据。与 OpenAI、Google 等厂商的安全报告相比,Anthropic 持续强调第三方评估的透明度,这条推文本身也是其负责任扩展策略的一部分。

Original · Claude

Fable 5.2 and Mythos 5.2 performed as well as, or better than, our recent models on most alignment measures. The model was also tested before release by external evaluators, including the @AISecurityInst , @FrontierDesign2 , and @METR_Evals .📊 300 ⚡

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Anthropic, Fable 5.2, Mythos 5.2, METR.

Source attribution: curated from Claude (@claudeai).