GPT-5.4 · HealthBench(48,000+ rubric items across 5,000 multi-turn conversations) 3311 characters · 附件证据(ev-gpt54-card-healthbench-length)

GPT-5.4 的 HealthBench · 48,000+ rubric items across 5,000 multi-turn conversations 官方出处。GPT-5.4 system-card evaluation · effort unspecified · 3311 characters。见《GPT-5.4 Thinking System Card | OpenAI》Section 3.6 “Health” / Table 6 “Health evaluations” / row “Average response length on HealthBench” / GPT-5.4 column。证据 ev-gpt54-card-healthbench-length。

原文与定位

GPT-5.4 · HealthBench(48,000+ rubric items across 5,000 multi-turn conversations),GPT-5.4 system-card evaluation · effort unspecified配置,Mean response length 3311 characters(具体成绩公布时间未确认)。见《GPT-5.4 Thinking System Card | OpenAI》Section 3.6 “Health” / Table 6 “Health evaluations” / row “Average response length on HealthBench” / GPT-5.4 column。 查看成绩条件

Average response length on HealthBench | GPT-5.4: 3311 chars
官方材料
GPT-5.4 Thinking System Card | OpenAI · PDF 第 10 页
定位
附件 · Section 3.6 “Health” / Table 6 “Health evaluations” / row “Average response length on HealthBench” / GPT-5.4 column
材料版本
Official GPT-5.4 Thinking System Card current saved edition; all 40 PDF pages and HTML were read, and rendered pages/figures were visually checked. The version publication date is not stated.
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

GPT-5.4 · HealthBench(48,000+ rubric items across 5,000 multi-turn conversations),GPT-5.4 system-card evaluation · effort unspecified配置,Mean response length 3311 characters(具体成绩公布时间未确认)。官方出处 GPT-5.4 Thinking System Card | OpenAI(附件,Official GPT-5.4 Thinking System Card current saved edition; all 40 PDF pages and HTML were read, and rendered pages/figures were visually checked. The version publication date is not stated.)Section 3.6 “Health” / Table 6 “Health evaluations” / row “Average response length on HealthBench” / GPT-5.4 column,https://deploymentsafety.openai.com/gpt-5-4-thinking/gpt-5-4-thinking.pdf#page=10。Benchlog https://benchlog.lekshome.cc/evidence/gpt-5-4/ev-gpt54-card-healthbench-length/

关联成绩与条件

GPT-5.4 · HealthBench 成绩与配置
指标与成绩配置与评测条件证据关系
Mean response length3311 characters原文 3311 chars · 单位 characters
公布 未确认
GPT-5.4 system-card evaluation · effort unspecified
设置
GPT-5.4 as named in system-card Table 6; reasoning effort and tool settings are not stated.
未知项
Reasoning effort、Tool access, grader details, and run settings
工具
工具条件未记录
推理预算
未记录
框架
未记录
协议
Mean response length on HealthBench, reported in characters.
条件
GPT-5.4, as named in the current saved system card; reasoning effort and evaluation run settings are not stated.

直接支持该成绩