Claude Haiku 5.5 · HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations · 成绩与配置

Claude Haiku 5.5 的 HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations 记录。本次发布有 5 条成绩记录,含 5 条精确值(可含历史记录),附 1 条来源定位。

成绩与评测条件

Claude Haiku 5.5 · HealthBench 成绩与配置
指标与成绩配置与评测条件出处
Length-adjusted score59.9%原文 59.9% · 单位 %
公布 未确认
Healthcare evaluations / low
设置
Safety classifiers enabled; low effort; run counts, scoring variants, and dates are recorded per result.
未知项
Healthcare evaluations use benchmark-specific run counts and scoring variants.
工具
禁止工具
推理预算
low
框架
未记录
协议
HealthBench length-adjusted score; Claude Opus 4.8 grader; no tools or customized system prompt.
条件
low effort; one run per level below max and five runs at max. Runs below max were made October 5; max is a five-run mean from October 2. Labeled figure value.
在排名中查看此成绩
Length-adjusted score60.2%原文 60.2% · 单位 %
公布 未确认
Healthcare evaluations / medium
设置
Safety classifiers enabled; medium effort; run counts, scoring variants, and dates are recorded per result.
未知项
Healthcare evaluations use benchmark-specific run counts and scoring variants.
工具
禁止工具
推理预算
medium
框架
未记录
协议
HealthBench length-adjusted score; Claude Opus 4.8 grader; no tools or customized system prompt.
条件
medium effort; one run per level below max and five runs at max. Runs below max were made October 5; max is a five-run mean from October 2. Labeled figure value.
在排名中查看此成绩
Length-adjusted score60.3%原文 60.3% · 单位 %
公布 未确认
Healthcare evaluations / high
设置
Safety classifiers enabled; high effort; run counts, scoring variants, and dates are recorded per result.
未知项
Healthcare evaluations use benchmark-specific run counts and scoring variants.
工具
禁止工具
推理预算
high
框架
未记录
协议
HealthBench length-adjusted score; Claude Opus 4.8 grader; no tools or customized system prompt.
条件
high effort; one run per level below max and five runs at max. Runs below max were made October 5; max is a five-run mean from October 2. Labeled figure value.
在排名中查看此成绩
Length-adjusted score61.1%原文 61.1% · 单位 %
公布 未确认
Healthcare evaluations / xhigh
设置
Safety classifiers enabled; xhigh effort; run counts, scoring variants, and dates are recorded per result.
未知项
Healthcare evaluations use benchmark-specific run counts and scoring variants.
工具
禁止工具
推理预算
xhigh
框架
未记录
协议
HealthBench length-adjusted score; Claude Opus 4.8 grader; no tools or customized system prompt.
条件
xhigh effort; one run per level below max and five runs at max. Runs below max were made October 5; max is a five-run mean from October 2. Labeled figure value.
在排名中查看此成绩
Length-adjusted score61.6%原文 61.6% · 单位 %
公布 未确认
Healthcare evaluations / max
设置
Safety classifiers enabled; max effort; run counts, scoring variants, and dates are recorded per result.
未知项
Healthcare evaluations use benchmark-specific run counts and scoring variants.
工具
禁止工具
推理预算
max
框架
未记录
协议
HealthBench length-adjusted score; Claude Opus 4.8 grader; no tools or customized system prompt.
条件
max effort; one run per level below max and five runs at max. Runs below max were made October 5; max is a five-run mean from October 2. Labeled figure value.
在排名中查看此成绩

出处与披露

测试版本与指标

别名 HealthBench

  • Length-adjusted score · % · 越高越好
  • Raw score · % · 越高越好
  • Mean response length · characters · 排序方向未确认
  • 长度校正分(0–100) · score · 越高越好
  • 未校正分(0–100) · score · 越高越好
  • 相对 GPT-5.4 的长度校正分差 · points · 排序方向未确认
  • Unadjusted score · % · 越高越好