Claude Opus 5 · OfficeQA(Pro 133-question subset) 66.9% · 附件证据(ev-op5-card-officeqa-pro)

Claude Opus 5 的 OfficeQA · Pro 133-question subset 官方出处。OfficeQA / OfficeQA Pro · Messages API · 66.9%。见《Claude Opus 5 System Card》System Card PDF p.180, §8.13.1 OfficeQA。证据 ev-op5-card-officeqa-pro。

原文与定位

Claude Opus 5 · OfficeQA(Pro 133-question subset),OfficeQA / OfficeQA Pro · Messages API配置,Exact-match accuracy 66.9%(具体成绩公布时间未确认)。见《Claude Opus 5 System Card》System Card PDF p.180, §8.13.1 OfficeQA。 查看成绩条件

Claude Opus 5 achieved 66.9% on OfficeQA Pro, the 133-question subset. The evaluation used extracted-text documents and code-execution tools on the public Messages API with production safeguards and Opus 4.8 fallback on classifier refusal.
官方材料
Claude Opus 5 System Card · PDF 第 180 页
定位
附件 · System Card PDF p.180, §8.13.1 OfficeQA
材料版本
Current saved PDF/TXT: 本地留存副本 and current-opus-5-system-card.txt; cover dated 2026-07-24; p.2 changelog records an update on 2026-08-19; later revision date, if any, unknown.
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 5 · OfficeQA(Pro 133-question subset),OfficeQA / OfficeQA Pro · Messages API配置,Exact-match accuracy 66.9%(具体成绩公布时间未确认)。官方出处 Claude Opus 5 System Card(附件,Current saved PDF/TXT: 本地留存副本 and current-opus-5-system-card.txt; cover dated 2026-07-24; p.2 changelog records an update on 2026-08-19; later revision date, if any, unknown.)System Card PDF p.180, §8.13.1 OfficeQA,https://www.anthropic.com/claude-opus-5-system-card#page=180。Benchlog https://benchlog.lekshome.cc/evidence/claude-opus-5/ev-op5-card-officeqa-pro/

关联成绩与条件

Claude Opus 5 · OfficeQA 成绩与配置
指标与成绩配置与评测条件证据关系
Exact-match accuracy66.9%原文 66.9% · 单位 %
公布 未确认
OfficeQA / OfficeQA Pro · Messages API
设置
Agentic evaluation with extracted-text documents in a sandbox and code-execution tools; public Messages API; production safety classifiers active with fallback to Opus 4.8 on refusal; card-wide default max effort applies.
未知项
The exact run count is not stated in this section.
工具
允许工具
推理预算
max
框架
未记录
协议
OfficeQA Pro; 133-question subset; agentic extracted-text evaluation with code-execution tools.
条件
Public Messages API; safety classifiers and Opus 4.8 fallback active; max effort by card-wide default.

直接支持该成绩

在排名中查看此成绩