Claude Opus 5.5 · Biomedical Image Analysis(System-card evaluation across public biomedical-image competition tasks; included competition set unspecified) 71.4% · 附件证据(ev-op55-card-biomedical-image)

Claude Opus 5.5 的 Biomedical Image Analysis · System-card evaluation across public biomedical-image competition tasks; included competition set unspecified 官方出处。System Card max effort (per-result benchmark harnesses) · 71.4%。见《Claude Opus 5.5 System Card》§8.17.7, pp.219-220, Figure 8.17.8.A。证据 ev-op55-card-biomedical-image。

原文与定位

Claude Opus 5.5 · Biomedical Image Analysis(System-card evaluation across public biomedical-image competition tasks; included competition set unspecified),System Card max effort (per-result benchmark harnesses)配置,Per-task normalized score 71.4%(具体成绩公布时间未确认)。见《Claude Opus 5.5 System Card》§8.17.7, pp.219-220, Figure 8.17.8.A。 查看成绩条件

Claude Opus 5.5 at 71.4%.
官方材料
Claude Opus 5.5 System Card · PDF 第 220 页
定位
附件 · §8.17.7, pp.219-220, Figure 8.17.8.A
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 · Biomedical Image Analysis(System-card evaluation across public biomedical-image competition tasks; included competition set unspecified),System Card max effort (per-result benchmark harnesses)配置,Per-task normalized score 71.4%(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)§8.17.7, pp.219-220, Figure 8.17.8.A,https://www.anthropic.com/claude-opus-5-5-system-card#page=220。Benchlog https://benchlog.lekshome.cc/evidence/claude-opus-5-5/ev-op55-card-biomedical-image/

关联成绩与条件

Claude Opus 5.5 · Biomedical Image Analysis 成绩与配置
指标与成绩配置与评测条件证据关系
Per-task normalized score71.4%原文 71.4% normalized score · 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
工具条件未记录
推理预算
max
框架
未记录
协议
Biomedical image analysis; 91 cases across 31 public competition tasks; one attempt per problem; no pretrained models or internet.
条件
Each task's metric is normalized baseline=0%, best published competition result=100%; use the proposed normalized-score metric, not accuracy. Source-wide conditions: 91 cases across 31 competition tasks; one attempt per problem; no pretrained models or internet. Each task metric is normalized so baseline=0% and best published competition result=100%; use normalized-score, not raw accuracy.

直接支持该成绩

在排名中查看此成绩