白泽（Whitzard）女娲实验室 ↗
NUWANUWA INDEX

搜索 NUWA Index

研究 · 测量协议 · 评测结果 · 观点

English
← Models

OpenAI

GPT-5.5

关于能力、行为倾向及可观察行动的已发表证据。

首次公开可用日期
2026-04-23
参数规模
未披露
元数据更新
2026-10-27
  • · 公开发布
  • · API 开放

评测档案记载 2026-04-24；该日期保留为来源记录，与独立维护的发布元数据区分。

来源 ↓

证据概述

按评估与协议整理已发表风险证据；证据缺失不代表风险较低。

NUWA 整理和解读这些来源。实验归属于原始执行方，本页不宣称 NUWA 已独立复现。

模型监测

代表性 Benchmark 覆盖

按协议和配置呈现结果，缺口标明待收录或待关联的证据。

Attempt to Persuade Eval (APE) →

2 项测量

  • 18%SaferAI · APE simulated persuadee; automated attempt evaluator; StrongREJECT refusal classifier
  • 22%SaferAI · APE simulated persuadee; automated attempt evaluator; StrongREJECT refusal classifier

全部已关联测量

各项结果保留原始协议与单位。

72 条观测

测量值与来源条件
系统／模型结果条件状态任务／条件评测方／报告日期
GPT-5.555.45%CyberGym · CyberGym pass@1条件不完整CyberGym Level 1 hard-task matched subsetCyberGym pass@12M tokens per task—Lyptus ResearchGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.5 · OpenAI AgentGPT-5.581.8%CyberGym · 成功率条件不完整level1score_10OpenAI AgentNot disclosedCyberGym official leaderboard and machine-data extraction发布 2026-04-23报告 —
GPT-5.588.12%CyberGym · CyberGym pass@1条件不完整CyberGym Level 1 hard-task matched subsetCyberGym pass@150M tokens per task—Lyptus ResearchGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.5 / Codex 0.133.0GPT-5.518.97%AgentCyberRange · 成功率（Pass@3）条件不完整Web exploitationlevel0CodexNot disclosedAgentCyberRange: Evaluate Agent Cyberattack Capabilities in Enterprise-like Cyber Ranges发布 2026-04-23报告 —
GPT-5.5 / Codex 0.147.0GPT-5.526.67%AgentCyberRange · 成功率（Pass@3）条件不完整Web exploitationlevel0CodexNot disclosedAgentCyberRange: Evaluate Agent Cyberattack Capabilities in Enterprise-like Cyber Ranges发布 2026-04-23报告 —
GPT-5.5 / Codex 0.133.0GPT-5.531.71%AgentCyberRange · 成功率（Pass@3）条件不完整Post exploitationlevel0CodexNot disclosedAgentCyberRange: Evaluate Agent Cyberattack Capabilities in Enterprise-like Cyber Ranges发布 2026-04-23报告 —
GPT-5.5 / Codex 0.147.0GPT-5.542.07%AgentCyberRange · 成功率（Pass@3）条件不完整Post exploitationlevel0CodexNot disclosedAgentCyberRange: Evaluate Agent Cyberattack Capabilities in Enterprise-like Cyber Ranges发布 2026-04-23报告 —
GPT-5.591.18%Cybench · Cybench pass@1条件不完整Cybench 34/40 task subsetCybench pass@120M total tokens; 100000 output/generationInspect EvalsSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.593.33%Cybench · Cybench category pass@1条件不完整Cybench CryptographyCybench category pass@1Inspect EvalsSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.583.33%Cybench · Cybench category pass@1条件不完整Cybench WebCybench category pass@1Inspect EvalsSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.5100%Cybench · Cybench category pass@1条件不完整Cybench PwnCybench category pass@1Inspect EvalsSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.5100%Cybench · Cybench category pass@1条件不完整Cybench Reverse Eng.Cybench category pass@1Inspect EvalsSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.5100%Cybench · Cybench category pass@1条件不完整Cybench ForensicsCybench category pass@1Inspect EvalsSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.550%Cybench · Cybench category pass@1条件不完整Cybench MiscCybench category pass@1Inspect EvalsSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.565.3%MASK · Continuations honesty95% Wilson CI [58.1, 72.0] (as reported)条件不完整MASK by archetype—Center for AI SafetyGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.569.6%MASK · Disinformation honesty95% Wilson CI [61.1, 77.0] (as reported)条件不完整MASK by archetype—Center for AI SafetyGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.563.3%MASK · Doubling-down honesty95% Wilson CI [54.4, 71.4] (as reported)条件不完整MASK by archetype—Center for AI SafetyGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.577%MASK · Known-facts honesty95% Wilson CI [70.9, 82.2] (as reported)条件不完整MASK by archetype—Center for AI SafetyGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.575.2%MASK · Provided-facts honesty95% Wilson CI [69.7, 79.9] (as reported)条件不完整MASK by archetype—Center for AI SafetyGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.587.5%MASK · Statistics honesty95% Wilson CI [79.4, 92.7] (as reported)条件不完整MASK by archetype—Center for AI SafetyGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighexplicit goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighexplicit goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighexplicit goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighlatent goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighlatent goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighlatent goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighambiguous goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighambiguous goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighambiguous goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighno goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighno goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Blackmail scenariohighno goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighexplicit goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighexplicit goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighexplicit goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighlatent goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighlatent goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighlatent goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighambiguous goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighambiguous goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighambiguous goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighno goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighno goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Leaking scenariohighno goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighexplicit goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighexplicit goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighexplicit goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighlatent goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighlatent goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighlatent goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighambiguous goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighambiguous goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighambiguous goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighno goal · ReplacementInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighno goal · RestrictionInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.50%Agentic misalignment · 有害行动发生率[0–11]条件不完整Agentic Misalignment Murder scenariohighno goal · No threatInspect Evals; GPT-5-mini transcript judgeSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.510%Agentic misalignment · Agentic Misalignment verbalized evaluation awareness条件不完整Agentic Misalignment evaluationAgentic Misalignment verbalized evaluation awareness—SaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.5 (xhigh)GPT-5.571%CTF-Archive-Diamond · Task accuracy条件不完整CTF-Archive-Diamond1M weighted tokensInspect ReActNIST / CAISICAISI Evaluation of DeepSeek V4 Pro发布 2026-04-23报告 2026-05-01
OpenAI GPT-5.5GPT-5.571%CTF-Archive-Diamond · benchmark_score条件不完整CTF-Archive-Diamondxhigh1M weighted tokensInspect built-in ReAct agentNIST / CAISICAISI Evaluation of DeepSeek V4 Pro发布 2026-04-23报告 2026-05-01
GPT-5.5 (xhigh)GPT-5.578%PortBench · Reported task accuracy条件不完整PortBench1M weighted tokensInspect ReActNIST / CAISICAISI Evaluation of DeepSeek V4 Pro发布 2026-04-23报告 2026-05-01
GPT-5.5 (OpenAI's default safety filters block all GPT-5.5 exploit attempts under default prompting.) · Codex CLIGPT-5.5129 countExploitGym · 指定目标成功数条件不完整v1on_target2h timeoutCodex CLINot disclosedExploitGym official leaderboard and benchmark page发布 2026-04-23报告 —
GPT-5.5 (OpenAI's default safety filters block all GPT-5.5 exploit attempts under default prompting.) · Codex CLIGPT-5.5208 countExploitGym · Flag 获取数条件不完整v1flag_captured2h timeoutCodex CLINot disclosedExploitGym official leaderboard and benchmark page发布 2026-04-23报告 —
GPT-5.558.4%SEC-bench Pro · Headline 成功率条件不完整260617 / overallheadlineHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=null; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.560.3%SEC-bench Pro · Completed 模式成功率条件不完整260617 / overallcompletedHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=null; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.547.6%SEC-bench Pro · Headline 成功率条件不完整260617 / v8headlineHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=["1.5h"]; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.548%SEC-bench Pro · Completed 模式成功率条件不完整260617 / v8completedHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=["1.5h"]; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.544.2%SEC-bench Pro · Headline 成功率条件不完整260617 / firefoxheadlineHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=["1.5h"]; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.544.7%SEC-bench Pro · Completed 模式成功率条件不完整260617 / firefoxcompletedHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=["1.5h"]; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.577.4%SEC-bench Pro · Headline 成功率条件不完整260617 / linuxheadlineHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=["1.5h"]; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.583.2%SEC-bench Pro · Completed 模式成功率条件不完整260617 / linuxcompletedHeadline UI: fixed budget; Completed UI: within timeout. Published summary timeout_caps=["1.5h"]; cap labels do not establish equal token or cost budgets.CodexNot disclosedSEC-bench Pro official leaderboard: selected 260617 and historical 260505发布 2026-04-23报告 —
GPT-5.518%Attempt to Persuade Eval (APE) · APE Turn-1 persuasion-attempt rate条件不完整APE Conspiracy (main persuade condition)highAPE simulated persuadee; automated attempt evaluator; StrongREJECT refusal classifierSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02
GPT-5.522%Attempt to Persuade Eval (APE) · APE Turn-1 persuasion-attempt rate条件不完整APE Undermining control (main persuade condition)highAPE simulated persuadee; automated attempt evaluator; StrongREJECT refusal classifierSaferAIGLM-5.2 Risk Evaluation Report发布 2026-04-23报告 2026-08-02

开发者披露

GPT-5.5 System Card →

关于GPT-5.5的开发者文档，核查范围和已关联协议见下方记录。

2026-04-23
仅核查章节结构

外部评估

按风险主题组织的证据

自主性 38 项协议结果

PortBench →

外部评测 · 执行方 美国 NIST／CAISI

Reported task accuracy
78%

见报告特定样本与分母。

欺骗 38 项协议结果

MASK by archetype →

引用结果 · 执行方 Center for AI Safety · 原始执行方 Center for AI Safety

Continuations honesty
65.3%
Disinformation honesty
69.6%
Doubling-down honesty
63.3%
Known-facts honesty
77.0%
Provided-facts honesty
75.2%
Statistics honesty
87.5%

见报告特定样本与分母。

网络安全 26 项协议结果

CTF-Archive-Diamond →

外部评测 · 执行方 美国 NIST／CAISI

Task accuracy
71%

见报告特定样本与分母。

Cybench →

外部评测 · 执行方 SaferAI

Cybench pass@1
91.18%

31/34

Cybench →

外部评测 · 执行方 SaferAI

Cybench category pass@1
93.33%

14/15

Cybench →

外部评测 · 执行方 SaferAI

Cybench category pass@1
83.33%

5/6

Cybench →

外部评测 · 执行方 SaferAI

Cybench category pass@1
100%

2/2

Cybench →

外部评测 · 执行方 SaferAI

Cybench category pass@1
100%

5/5

Cybench →

外部评测 · 执行方 SaferAI

Cybench category pass@1
100%

4/4

Cybench →

外部评测 · 执行方 SaferAI

Cybench category pass@1
50%

1/2

CTF-Archive-Diamond →

外部评测 · 执行方 美国 NIST／CAISI

benchmark_score
71%

原文未披露本结果的样本量

化生放核 未关联结构化结果

目录缺口不代表低风险。尚未提取的发现及协议范围请见关联报告。

虚假信息与影响 2 项协议结果

APE Turn-1 attempts →

外部评测 · 执行方 SaferAI

APE Turn-1 persuasion-attempt rate
18%

来源分母：100（单位未披露）

APE Turn-1 attempts →

外部评测 · 执行方 SaferAI

APE Turn-1 persuasion-attempt rate
22%

来源分母：100（单位未披露）

自主物理危害 未关联结构化结果

目录缺口不代表低风险。尚未提取的发现及协议范围请见关联报告。

来源核查与证据缺口

CyberGym →

整理 CyberGym 官网报告的评测结果，保留受测系统和运行配置。

官网收录不同配置下的系统提交。本结果的任务数、预算和执行日期未完整披露；未定义指标及身份待核实行已排除。JSON 原始比例按官网展示规则乘以 100 并保留一位小数，转换为百分比。

ExploitGym →

整理 ExploitGym 官网报告的评测结果，保留受测系统和运行配置。

指定目标成功与 Flag 获取是不同的计数。任务版本和超时截点分别保留，不将目录总任务数视为各提交的分母。

SEC-bench Pro →

整理 SEC-bench Pro 官网报告的评测结果，保留受测系统和运行配置。

Headline 与 Completed 使用不同纳入规则，分子与分母都可能变化。模型提供方及服务后端不等于实际评测方。

AgentCyberRange →

整理 AgentCyberRange 官网报告的评测结果，保留受测系统和运行配置。

Web 利用与后渗透分别评测。Pass@3 保留官网定义；各行的预算、样本量和执行日期未披露。

GLM-5.2 Risk Evaluation Report, Version 1.1 →

对 SaferAI · GLM-5.2 的逐来源提取与核查，发现和数值均保留原始归属。

来源转录不等于实验复现。未确定的日期、执行方及配置保留为未知；原文分组不作为经核验排名。

CAISI Evaluation of DeepSeek V4 Pro →

对 CAISI · DeepSeek V4 Pro 的逐来源提取与核查，发现和数值均保留原始归属。

来源转录不等于实验复现。未确定的日期、执行方及配置保留为未知；原文分组不作为经核验排名。

缺失分数可能来自尚未提取的章节、未披露配置、身份待核实或尚无实验。这些问题需要分别处理，均不代表零分。

解读与开放问题

报告差异可能来自任务选择、模型端点、提示、预算、判别器和发表时间。证据依据实际协议分组，不推断模型整体安全等级。

参考文献

  1. [1] OpenAI — introducing gpt 5 5 ; Publication date and availability section; Parameter counts not disclosed in launch article
  2. [2] GLM-5.2 Risk Evaluation Report. 2026-08-02. v1.1 · revised 26 August 2026 ; Appendix A.7; Table A.3, pp. 41–42; p.6 Fig.1; p.6 Fig.2 printed bar label
  3. [3] CAISI Evaluation of DeepSeek V4 Pro. 2026-05-01. 1 May 2026 ; Capability Results; Figure 2; *** footnote; Capability Results; Figure 2; Figure 2 HTML table, CTF-Archive-Diamond, OpenAI GPT-5.5 (xhigh)
  4. [4] GPT-5.5 System Card. 2026-04-23
  5. [5] CyberGym official leaderboard and machine-data extraction ; /level1/51/score_10
  6. [6] ExploitGym official leaderboard and benchmark page ; /results/3/on_target; /results/3/flag_captured
  7. [7] SEC-bench Pro official leaderboard: selected 260617 and historical 260505 ; /0/results/0/score_modes/headline/score; /0/results/0/score_modes/completed/score; /1/results/0/score_modes/headline/score
  8. [8] AgentCyberRange: Evaluate Agent Cyberattack Capabilities in Enterprise-like Cyber Ranges ; routes-B-0Mubrq.js::QR/1/level0/pass3; routes-B-0Mubrq.js::QR/2/level0/pass3; routes-B-0Mubrq.js::$R/1/level0/pass3
  9. [9] Jack Payne, Sean Peters. GPT-5.5 Saturates Our Offensive Cybersecurity Time Horizons. 2026-05-27