Claude Opus 4.6 · Real-World Finance(Anthropic internal evaluation; about 50 analyst-workflow tasks; version unspecified) 64.1% · 附件证据(ev-op46-card-real-world-finance)

Claude Opus 4.6 的 Real-World Finance · Anthropic internal evaluation; about 50 analyst-workflow tasks; version unspecified 官方出处。Real-World Finance · Anthropic internal · 64.1%。见《Claude Opus 4.6 System Card》PDF pp.27–28, §2.14.4 and Figure 2.14.4.A。证据 ev-op46-card-real-world-finance。

原文与定位

Claude Opus 4.6 · Real-World Finance(Anthropic internal evaluation; about 50 analyst-workflow tasks; version unspecified),Real-World Finance · Anthropic internal配置,Percentage task completion 64.1%(具体成绩公布时间未确认)。见《Claude Opus 4.6 System Card》PDF pp.27–28, §2.14.4 and Figure 2.14.4.A。 查看成绩条件

The chart reports 64.1% average task-completion score for Claude Opus 4.6. Anthropic describes an internal evaluation of about 50 analyst-workflow tasks, graded by rubrics/preferences; the card notes no independent third-party validation.
官方材料
Claude Opus 4.6 System Card · PDF 第 28 页
定位
附件 · PDF pp.27–28, §2.14.4 and Figure 2.14.4.A
材料版本
Current PDF/TXT; local source 本地留存副本 and 本地留存副本; 213 pages; changelog through 2026-03-06; full text and visuals reviewed.
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 Opus 4.6 · Real-World Finance(Anthropic internal evaluation; about 50 analyst-workflow tasks; version unspecified),Real-World Finance · Anthropic internal配置,Percentage task completion 64.1%(具体成绩公布时间未确认)。官方出处 Claude Opus 4.6 System Card(附件,Current PDF/TXT; local source 本地留存副本 and 本地留存副本; 213 pages; changelog through 2026-03-06; full text and visuals reviewed.)PDF pp.27–28, §2.14.4 and Figure 2.14.4.A,https://www.anthropic.com/claude-opus-4-6-system-card#page=28。Benchlog https://benchlog.lekshome.cc/evidence/claude-opus-4-6/ev-op46-card-real-world-finance/

关联成绩与条件

Claude Opus 4.6 · Real-World Finance 成绩与配置
指标与成绩配置与评测条件证据关系
Percentage task completion64.1%原文 64.1% · 单位 %
公布 未确认
Real-World Finance · Anthropic internal
设置
Anthropic 内部金融工作流评测;约 50 项真实分析师任务;code execution 与工具使用 agentic harness;effort 未说明。
未知项
内部评测,未独立第三方验证;PDF 未给 trial 数。
工具
允许工具
推理预算
未记录
框架
未记录
协议
Anthropic Real-World Finance;overall percentage task completion。
条件
约50项内部分析师工作流任务;code execution/tool-use agentic harness;rubric/preferences评 finance knowledge、completeness/accuracy、presentation;无独立第三方验证。effort 与 trial 数未说明。

直接支持该成绩

在排名中查看此成绩