Claude Opus 4.7 · DeepSearchQA(900 prompts;单代理与多代理按协议区分) 89.1% · 附件证据(ev-op47-card-deepsearchqa)

Claude Opus 4.7 的 DeepSearchQA · 900 prompts;单代理与多代理按协议区分 官方出处。DeepSearchQA,single-agent · 89.1%。见《Claude Opus 4.7 System Card》§8.8.3, Table 8.8.3.B and Figure 8.8.3.A, PDF pp.199-200。证据 ev-op47-card-deepsearchqa。

原文与定位

Claude Opus 4.7 · DeepSearchQA(900 prompts;单代理与多代理按协议区分),DeepSearchQA,single-agent配置,F1 score 89.1%(具体成绩公布时间未确认)。见《Claude Opus 4.7 System Card》§8.8.3, Table 8.8.3.B and Figure 8.8.3.A, PDF pp.199-200。 查看成绩条件

Claude Opus 4.7: F1 89.1% ±1.8%; 80.7% fully correct, 7.0% fully incorrect, 3.9% correct with excessive answers. The figure uses up to 10M total tokens and 200k compaction trigger.
官方材料
Claude Opus 4.7 System Card · PDF 第 200 页
定位
附件 · §8.8.3, Table 8.8.3.B and Figure 8.8.3.A, PDF pp.199-200
材料版本
Current saved PDF/TXT: 本地留存副本 and current-opus-4-7-system-card.txt; 232 PDF pages; metadata created/modified 2026-04-16 22:11:48 China Standard Time. Full current-card review completed 2026-11-16; this file is not established as launch-day bytes.
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 4.7 · DeepSearchQA(900 prompts;单代理与多代理按协议区分),DeepSearchQA,single-agent配置,F1 score 89.1%(具体成绩公布时间未确认)。官方出处 Claude Opus 4.7 System Card(附件,Current saved PDF/TXT: 本地留存副本 and current-opus-4-7-system-card.txt; 232 PDF pages; metadata created/modified 2026-04-16 22:11:48 China Standard Time. Full current-card review completed 2026-11-16; this file is not established as launch-day bytes.)§8.8.3, Table 8.8.3.B and Figure 8.8.3.A, PDF pp.199-200,https://www.anthropic.com/claude-opus-4-7-system-card#page=200。Benchlog https://benchlog.lekshome.cc/evidence/claude-opus-4-7/ev-op47-card-deepsearchqa/

关联成绩与条件

Claude Opus 4.7 · DeepSearchQA 成绩与配置
指标与成绩配置与评测条件证据关系
F1 score89.1%原文 89.1% ±1.8% · 单位 %
公布 未确认
DeepSearchQA,single-agent
设置
900 prompts; web search, web fetch, programmatic tool calling; adaptive thinking, max effort; context compaction at 200K; up to 10M total tokens.
未知项
未另记
工具
允许工具
推理预算
max
框架
未记录
协议
DeepSearchQA; 900 prompts; single-agent; F1.
条件
Web search, web fetch, programmatic tool calling; adaptive thinking and max effort; 200K compaction trigger; up to 10M total tokens. Table 8.8.3.B reports ±1.8%.

直接支持该成绩

在排名中查看此成绩