Gemini 3 Pro · Humanity's Last Exam(2,500-question benchmark) 37.5% · 附件证据(eval-hle-no-tools)

Gemini 3 Pro 的 Humanity's Last Exam · 2,500-question benchmark 官方出处。Gemini 3 Pro API default · 37.5%。见《Gemini 3 Pro - Evaluations Approach, Methodology & Results》PDF p.3, results table, Humanity's Last Exam / No tools row, Gemini 3 Pro column。证据 eval-hle-no-tools。

原文与定位

Gemini 3 Pro · Humanity's Last Exam(2,500-question benchmark),Gemini 3 Pro API default配置,Accuracy 37.5%(具体成绩公布时间未确认)。见《Gemini 3 Pro - Evaluations Approach, Methodology & Results》PDF p.3, results table, Humanity's Last Exam / No tools row, Gemini 3 Pro column。 查看成绩条件

Humanity's Last Exam / No tools / Gemini 3 Pro: 37.5%
官方材料
Gemini 3 Pro - Evaluations Approach, Methodology & Results · PDF 第 3 页
定位
附件 · PDF p.3, results table, Humanity's Last Exam / No tools row, Gemini 3 Pro column
材料版本
3-page evaluation PDF; results as of November 2025
版本公布
未确认
披露时间
· 已有首发依据
采集时间
历史依据
已实际取得首发当天官方PDF历史回放 https://web.archive.org/web/20251118163819id_/https://storage.googleapis.com/deepmind-media/gemini/gemini_3_pro_model_evaluation.pdf;回放原件与已完整核读的当前PDF完全相同。原件及响应记录:本轮整合目录 gemini-3-evaluation-historical-cdn.pdf。
引用文本

Gemini 3 Pro · Humanity's Last Exam(2,500-question benchmark),Gemini 3 Pro API default配置,Accuracy 37.5%(具体成绩公布时间未确认)。官方出处 Gemini 3 Pro - Evaluations Approach, Methodology & Results(附件,3-page evaluation PDF; results as of November 2025)PDF p.3, results table, Humanity's Last Exam / No tools row, Gemini 3 Pro column,https://deepmind.google/models/evals-methodology/gemini-3-pro#page=3。Benchlog https://benchlog.lekshome.cc/evidence/gemini-3-pro/eval-hle-no-tools/

关联成绩与条件

Gemini 3 Pro · 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
Accuracy37.5%原文 37.5% · 单位 %
公布 未确认
Gemini 3 Pro API default
设置
Gemini API model id gemini-3-pro-preview; default sampling settings. Gemini results are pass@1 and single-attempt unless specified; smaller benchmarks average multiple trials.
未知项
The evaluation table does not specify a thinking budget or inference-time compute setting.
工具
禁止工具
推理预算
未记录
框架
未记录
协议
No tools; academic reasoning full set.
条件
Gemini API model gemini-3-pro-preview; default sampling settings; pass@1 unless stated otherwise.

直接支持该成绩

在排名中查看此成绩