Claude Opus 5.5 · OfficeQA(Full set) 78.9% · 附件证据(ev-op55-card-officeqa)

Claude Opus 5.5 的 OfficeQA · Full set 官方出处。System Card max effort (per-result benchmark harnesses) · 78.9%。见《Claude Opus 5.5 System Card》§8.14.1, p.208。证据 ev-op55-card-officeqa。

原文与定位

Claude Opus 5.5 · OfficeQA(Full set),System Card max effort (per-result benchmark harnesses)配置,Exact-match accuracy 78.9%(具体成绩公布时间未确认)。见《Claude Opus 5.5 System Card》§8.14.1, p.208。 查看成绩条件

Claude Opus 5.5 achieved 78.9% on OfficeQA and 67.7% on OfficeQA Pro (max effort, mean of five runs).
官方材料
Claude Opus 5.5 System Card · PDF 第 208 页
定位
附件 · §8.14.1, p.208
材料版本
官方当前 System Card;封面日期 2026-09-22;本地 PDF/TXT:本地留存副本、current-opus-5-5-system-card.txt;230 页全文核读;当前修订时间未知
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 5.5 · OfficeQA(Full set),System Card max effort (per-result benchmark harnesses)配置,Exact-match accuracy 78.9%(具体成绩公布时间未确认)。官方出处 Claude Opus 5.5 System Card(附件,官方当前 System Card;封面日期 2026-09-22;本地 PDF/TXT:本地留存副本、current-opus-5-5-system-card.txt;230 页全文核读;当前修订时间未知)§8.14.1, p.208,https://www.anthropic.com/claude-opus-5-5-system-card#page=208。Benchlog https://benchlog.lekshome.cc/evidence/claude-opus-5-5/ev-op55-card-officeqa/

关联成绩与条件

Claude Opus 5.5 · OfficeQA 成绩与配置
指标与成绩配置与评测条件证据关系
Exact-match accuracy78.9%原文 78.9% · 单位 %
公布 未确认
System Card max effort (per-result benchmark harnesses)
设置
System Card §8.1 default: adaptive thinking/max effort and default sampling settings unless a section says otherwise. Benchmark tools, evaluator, safeguards, fallback, and repetitions remain per-result conditions; this config represents the stated model effort, not harness equivalence.
未知项
Exact run time is not disclosed.、Per-benchmark tools, safeguards, and repetition counts are specified only where the source states them.
工具
允许工具
推理预算
max
框架
未记录
协议
OfficeQA full set; max effort; mean over five runs; extracted Treasury Bulletin text in sandbox with code execution.
条件
Exact full-set task count is not stated in the card. Source-wide conditions: Extracted-text corpus in a sandbox; code execution tools; OfficeQA Pro is the 133-question subset.

直接支持该成绩

在排名中查看此成绩