Official release article with full visible benchmark appendix; revision date unknown; Original page date label 2025-12-11; current page revision date is unknown.
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-5.2 · Humanity's Last Exam(2,500-question benchmark),GPT-5.2 Thinking (xhigh)配置,Accuracy 34.5%(具体成绩公布时间未确认);GPT-5.2 · Humanity's Last Exam(2,500-question benchmark),GPT-5.2 Pro (xhigh)配置,Accuracy 36.6%(具体成绩公布时间未确认);GPT-5.2 · Humanity's Last Exam(2,500-question benchmark),GPT-5.2 Thinking (xhigh)配置,Accuracy 45.5%(具体成绩公布时间未确认);GPT-5.2 · Humanity's Last Exam(2,500-question benchmark),GPT-5.2 Pro (xhigh)配置,Accuracy 50%(具体成绩公布时间未确认)。官方出处 隆重推出 GPT-5.2 | OpenAI(主文,Official release article with full visible benchmark appendix; revision date unknown; Original page date label 2025-12-11; current page revision date is unknown.)Appendix / Academic / HLE rows,https://openai.com/zh-Hans-CN/index/introducing-gpt-5-2/。Benchlog https://benchlog.lekshome.cc/evidence/gpt-5-2/ev-gpt52-hle/
关联成绩与条件
GPT-5.2 · Humanity's Last Exam 成绩与配置
指标与成绩
配置与评测条件
证据关系
Accuracy34.5%原文 34.5% · 单位 %
公布 未确认
GPT-5.2 Thinking (xhigh)
设置
Highest available API reasoning effort xhigh, except professional benchmark tasks.
未知项
未另记
工具
禁止工具
推理预算
xhigh
框架
未记录
协议
HLE, no tools; exact subset/version not restated in table.
GPT-5.2 · Humanity's Last Exam (2,500-question benchmark) · Main text evidence (ev-gpt52-hle)
Official source for Humanity's Last Exam · 2,500-question benchmark in GPT-5.2. See “隆重推出 GPT-5.2 | OpenAI” at Appendix / Academic / HLE rows. Evidence ev-gpt52-hle.
Official release article with full visible benchmark appendix; revision date unknown; Original page date label 2025-12-11; current page revision date is unknown.
Version published
Unconfirmed
Disclosure date
Unconfirmed · Launch or later disclosure unconfirmed
Collected
Historical basis
No reliable historical version evidence available
Citation text
GPT-5.2 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.2 Thinking (xhigh), Accuracy 34.5% (exact score publication date unconfirmed); GPT-5.2 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.2 Pro (xhigh), Accuracy 36.6% (exact score publication date unconfirmed); GPT-5.2 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.2 Thinking (xhigh), Accuracy 45.5% (exact score publication date unconfirmed); GPT-5.2 · Humanity's Last Exam (2,500-question benchmark), configuration GPT-5.2 Pro (xhigh), Accuracy 50% (exact score publication date unconfirmed). Official source: 隆重推出 GPT-5.2 | OpenAI (Main text, Official release article with full visible benchmark appendix; revision date unknown; Original page date label 2025-12-11; current page revision date is unknown.), Appendix / Academic / HLE rows, https://openai.com/zh-Hans-CN/index/introducing-gpt-5-2/. Benchlog https://benchlog.lekshome.cc/evidence/gpt-5-2/ev-gpt52-hle/?lang=en
Related scores and conditions
GPT-5.2 · Humanity's Last Exam scores and configurations
Metric and score
Configuration and evaluation conditions
Evidence relation
Accuracy34.5%Reported 34.5% · Unit %
Published Unconfirmed
GPT-5.2 Thinking (xhigh)
Settings
Highest available API reasoning effort xhigh, except professional benchmark tasks.
Unknowns
none recorded
Tools
Tools forbidden
Reasoning budget
xhigh
Framework
not recorded
Protocol
HLE, no tools; exact subset/version not restated in table.