GPT-5 mini · SWE-Lancer(IC Diamond set) 51% · 附件证据(ev-gpt5-mini-swelancer-ic-diamond-pass1-current60)

GPT-5 mini 的 SWE-Lancer · IC Diamond set 官方出处。Preparedness · non-public verbosity · 51%。见《GPT-5 System Card (current 60-page revision)》Current 60-page PDF pp.39-40, §5.1.3.4, Figure 24; archived pp.38-39, §5.1.2.7.。证据 ev-gpt5-mini-swelancer-ic-diamond-pass1-current60。

原文与定位

GPT-5 mini · SWE-Lancer(IC Diamond set),Preparedness · non-public verbosity配置,Score 51%(成绩公布 2025-08-07)。见《GPT-5 System Card (current 60-page revision)》Current 60-page PDF pp.39-40, §5.1.3.4, Figure 24; archived pp.38-39, §5.1.2.7.。 查看成绩条件

Figure 24, SWE-Lancer IC SWE (preceding section identifies the IC SWE Diamond set), no browsing: gpt-5-thinking-mini 51%.
官方材料
GPT-5 System Card (current 60-page revision) · PDF 第 40 页
定位
附件 · Current 60-page PDF pp.39-40, §5.1.3.4, Figure 24; archived pp.38-39, §5.1.2.7.
材料版本
Current canonical PDF dated August 13, 2025; text pp.1-60 fully read; all 45 table/figure pages visually inspected, plus p.31; detailed scope in 2026-11-16-review-gpt-early/PROGRESS.md. Exact current revision date 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 mini · SWE-Lancer(IC Diamond set),Preparedness · non-public verbosity配置,Score 51%(成绩公布 2025-08-07)。官方出处 GPT-5 System Card (current 60-page revision)(附件,Current canonical PDF dated August 13, 2025; text pp.1-60 fully read; all 45 table/figure pages visually inspected, plus p.31; detailed scope in 2026-11-16-review-gpt-early/PROGRESS.md. Exact current revision date unknown.)Current 60-page PDF pp.39-40, §5.1.3.4, Figure 24; archived pp.38-39, §5.1.2.7.,https://cdn.openai.com/gpt-5-system-card.pdf#page=40。Benchlog https://benchlog.lekshome.cc/evidence/gpt-5-mini/ev-gpt5-mini-swelancer-ic-diamond-pass1-current60/

关联成绩与条件

GPT-5 mini · SWE-Lancer 成绩与配置
指标与成绩配置与评测条件证据关系
Score51%原文 51% · 单位 %
公布
Preparedness · non-public verbosity
设置
High reasoning effort; maximum trained-in verbosity higher than public API high; one attempt per problem; no browsing. GPT-5 System Card current p.37 states that all Preparedness evaluations use this non-public verbosity.
未知项
该verbosity未通过当时公共API提供;不映射为普通High。
工具
工具条件未记录
推理预算
High
框架
未记录
协议
SWE-Lancer IC SWE Diamond, pass@1; no browsing.
条件
High reasoning effort; one attempt per problem; launch-candidate figure label. Card explicitly states all Preparedness evaluations use maximum trained-in verbosity above the API high setting; retained raw, excluded from ordinary presentation.

直接支持该成绩