Claude Opus · HealthBench Professional · 525 physician-authored conversations · 历代记录

按发布时间查看 Claude Opus 的 3 次发布在 HealthBench Professional · 525 physician-authored conversations 的成绩、配置与披露。

仅为直接前代间的唯一、有效精确值计算差值,要求单位、指标与明确记录的配置和条件一致。

Claude Opus 4.8

· 本次发布有 1 条成绩记录,含 1 条精确值(可含历史记录)

Claude Opus 4.8 · HealthBench Professional 成绩与配置
指标与成绩配置与评测条件出处
Length-adjusted score55.8%原文 55.8% · 单位 %
公布 未确认
HealthBench Professional
设置
525 physician-authored conversations;adaptive thinking/max effort;Sonnet 4.6 grader;平均 5 trials;论文定义的 length adjustment。
未知项
未另记
工具
工具条件未记录
推理预算
max
框架
未记录
协议
HealthBench Professional, 525 physician-authored conversations; LLM-as-a-Judge; length-adjusted score per original paper.
条件
Adaptive thinking/max effort; Claude Sonnet 4.6 grader; average over 5 trials; Figure 8.14.A shows 95% CI.
在排名中查看此成绩

Claude Opus 5

· 本次发布有 2 条成绩记录,含 2 条精确值(可含历史记录)

与直接前代的记录缺少满足差值计算的同条件依据。

Claude Opus 5 · HealthBench Professional 成绩与配置
指标与成绩配置与评测条件出处
Length-adjusted score59.8%原文 59.8% · 单位 %
公布 未确认
HealthBench Professional · max
设置
Max effort; no tools or customized system prompts; Opus 4.8 grader; five-trial average; length-adjustment method from the HealthBench Professional paper.
未知项
未另记
工具
禁止工具
推理预算
max
框架
未记录
协议
HealthBench Professional; length-adjusted score.
条件
59.8%; card also reports raw score 73.4%; max effort, no tools/custom system prompt, Opus 4.8 grader, five-trial average.
在排名中查看此成绩
Raw score73.4%原文 73.4% · 单位 %
公布 未确认
HealthBench Professional · max
设置
Max effort; no tools or customized system prompts; Opus 4.8 grader; five-trial average; length-adjustment method from the HealthBench Professional paper.
未知项
未另记
工具
禁止工具
推理预算
max
框架
未记录
协议
HealthBench Professional; raw rubric score.
条件
Max effort; no tools or customized system prompt; Opus 4.8 grader; five-trial average.
在排名中查看此成绩

Claude Opus 5.5

· 本次发布有 2 条成绩记录,含 2 条精确值(可含历史记录)

与直接前代的记录缺少满足差值计算的同条件依据。

Claude Opus 5.5 · HealthBench Professional 成绩与配置
指标与成绩配置与评测条件出处
Raw score77.1%原文 77.1% raw rubric score · 单位 %
公布 未确认
HealthBench Professional max effort with safeguards
设置
Adaptive thinking/max effort; safety classifiers enabled and refusal fallback to Opus 5; no tools or customized system prompts; five trials.
未知项
未另记
工具
禁止工具
推理预算
max
框架
未记录
协议
HealthBench Professional; raw score; five trials; no tools or customized system prompt.
条件
Safety classifiers enabled with refusal fallback to Opus 5. Source-wide conditions: Five trials; no tools or customized system prompts; safety classifiers enabled with refusal fallback to Opus 5.
在排名中查看此成绩
Length-adjusted score65.6%原文 65.6% length-adjusted score · 单位 %
公布 未确认
HealthBench Professional max effort with safeguards
设置
Adaptive thinking/max effort; safety classifiers enabled and refusal fallback to Opus 5; no tools or customized system prompts; five trials.
未知项
未另记
工具
禁止工具
推理预算
max
框架
未记录
协议
HealthBench Professional; length-adjusted score; five trials; no tools or customized system prompt.
条件
Length adjustment penalizes verbose responses; safety classifiers enabled with refusal fallback to Opus 5. Source-wide conditions: Five trials; no tools or customized system prompts; safety classifiers enabled with refusal fallback to Opus 5.
在排名中查看此成绩