Claude Fable 5.1 · FrontierCode(v1.1 Main set) 50.9% · 附件证据(ev-f51-card-frontiercode-main)

Claude Fable 5.1 的 FrontierCode · v1.1 Main set 官方出处。Medium effort · 50.9%。见《Claude Fable 5.1 and Claude Mythos 5.1 System Card》本地留存副本, PDF p.169。证据 ev-f51-card-frontiercode-main。

原文与定位

Claude Fable 5.1 · FrontierCode(v1.1 Main set),Medium effort配置,Overall score 50.9%(具体成绩公布时间未确认)。见《Claude Fable 5.1 and Claude Mythos 5.1 System Card》本地留存副本, PDF p.169。 查看成绩条件

On FrontierCode 1.1 (Main), Fable 5.1 scored 50.9% at medium effort; the same page separately reports 63.6% on Extended.
官方材料
Claude Fable 5.1 and Claude Mythos 5.1 System Card · PDF 第 169 页
定位
附件 · 本地留存副本, PDF p.169
材料版本
PDF dated September 1, 2026; 212 pages. Full text review completed: pp.1-212. All 64 low-text or figure pages were rendered and visually inspected in 16 sheets; no additional standard Fable 5.1 capability scores were found in the supplementary safety/alignment sections. Revision date unknown. Local: 本地留存副本 and .txt.
版本公布
未确认
披露时间
未确认 · 首发/后补未确认
采集时间
历史依据
暂无可靠历史版本依据
引用文本

Claude Fable 5.1 · FrontierCode(v1.1 Main set),Medium effort配置,Overall score 50.9%(具体成绩公布时间未确认)。官方出处 Claude Fable 5.1 and Claude Mythos 5.1 System Card(附件,PDF dated September 1, 2026; 212 pages. Full text review completed: pp.1-212. All 64 low-text or figure pages were rendered and visually inspected in 16 sheets; no additional standard Fable 5.1 capability scores were found in the supplementary safety/alignment sections. Revision date unknown. Local: 本地留存副本 and .txt.)本地留存副本, PDF p.169,https://www.anthropic.com/claude-fable-5-1-mythos-5-1-system-card#page=169。Benchlog https://benchlog.lekshome.cc/evidence/claude-fable-and-mythos-5-1/ev-f51-card-frontiercode-main/

关联成绩与条件

Claude Fable 5.1 · FrontierCode 成绩与配置
指标与成绩配置与评测条件证据关系
Overall score50.9%原文 50.9% · 单位 %
公布 未确认
Medium effort
设置
Claude Fable 5.1; adaptive thinking at medium effort.
未知项
Sampling parameters are not specified for each chart point.
工具
工具条件未记录
推理预算
medium
框架
未记录
协议
FrontierCode v1.1 Main; medium effort.
条件
Composite of held-out functional tests and weighted code-quality rubric.

直接支持该成绩

在排名中查看此成绩