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
← NUWA Index

来源档案

ExploitGym

整理 ExploitGym 官网报告的评测结果，保留受测系统和运行配置。

原始来源与出处 ↓

评测结果

数据报告方: ExploitGym website · 实际执行方: 各提交的执行方未完整确定 · 抓取时间: 2026-10-26

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。

图表按来源顺序展示 8／8 个系统

指定目标成功数 · v1

ExploitGym指定目标成功数 · v1来源顺序 · 原始单位 · 保留配置差异082164246328GPT-5.6 Sol (reasoning max) · Codex CLI: 293 count; v1; 6h timeoutGPT-5.6 Sol (reasoning max) · CodexCLI293 countGPT-5.6 Sol (reasoning max) · Codex CLI: 216 count; v1; 2h timeout cutoffGPT-5.6 Sol (reasoning max) · CodexCLI216 countMuse Spark 1.1 (helpful-only versoin) · Meta Agent (off-the-shelf agent harness): 7 count; v1; 4h timeoutMuse Spark 1.1 (helpful-only versoin)· Meta Agent (off-the-shelf agentharness)7 countGPT-5.5 (OpenAI's default safety filters block all GPT-5.5 exploit attempts under default prompting.) · Codex CLI: 129 count; v1; 2h timeoutGPT-5.5 (OpenAI's default safetyfilters block all GPT-5.5 exploitattempts under default prompting.) ·Codex CLI129 countGPT-5.4 · Codex CLI: 61 count; v1; 2h timeoutGPT-5.4 · Codex CLI61 countClaude Opus 4.6 (Results obtained in collaboration with Anthropic.) · Claude Code: 16 count; v1; 2h timeoutClaude Opus 4.6 (Results obtained incollaboration with Anthropic.) ·Claude Code16 countClaude Opus 4.7 · Claude Code: 12 count; v1; 2h timeoutClaude Opus 4.7 · Claude Code12 countGLM-5.1 · Claude Code: 4 count; v1; 2h timeoutGLM-5.1 · Claude Code4 counthttps://www.cybergym.io/exploitgym/
  1. Codex CLI · 6h timeout
  2. Codex CLI · 2h timeout cutoff
  3. Meta Agent (off-the-shelf agent harness) · 4h timeout
  4. Codex CLI · 2h timeout
  5. Codex CLI · 2h timeout
  6. Claude Code · 2h timeout
  7. Claude Code · 2h timeout
  8. Claude Code · 2h timeout
来源报告的系统及评测条件
系统结果任务集／模式运行配置日期／归属
GPT-5.6 Sol (reasoning max) · Codex CLI293 countv1on_targetCodex CLI6h timeout · ? 次尝试2026-07-13来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GPT-5.6 Sol (reasoning max) · Codex CLI216 countv1on_targetCodex CLI2h timeout cutoff · ? 次尝试未注明来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Muse Spark 1.1 (helpful-only versoin) · Meta Agent (off-the-shelf agent harness)7 countv1on_targetMeta Agent (off-the-shelf agent harness)4h timeout · ? 次尝试2026-07-09来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GPT-5.5 (OpenAI's default safety filters block all GPT-5.5 exploit attempts under default prompting.) · Codex CLI129 countv1on_targetCodex CLI2h timeout · ? 次尝试2026-06-16来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GPT-5.5 (OpenAI's default safety filters block all GPT-5.5 exploit attempts under default prompting.) · Codex CLI208 countv1flag_capturedCodex CLI2h timeout · ? 次尝试2026-06-16来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GPT-5.4 · Codex CLI61 countv1on_targetCodex CLI2h timeout · ? 次尝试2026-06-16来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Opus 4.6 (Results obtained in collaboration with Anthropic.) · Claude Code16 countv1on_targetClaude Code2h timeout · ? 次尝试2026-06-16来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Gemini 3.1 Pro · Gemini CLI12 countv0on_targetGemini CLI2h timeout · ? 次尝试2026-05-13来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Opus 4.7 · Claude Code12 countv1on_targetClaude Code2h timeout · ? 次尝试2026-06-16来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GLM-5.1 · Claude Code4 countv1on_targetClaude Code2h timeout · ? 次尝试2026-06-16来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Mythos 5247 countVersion not specifiedon_target未注明6h timeout · ? 次尝试2026-06-09来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Mythos 5181 countVersion not specifiedon_target未注明2h timeout cutoff · ? 次尝试未注明来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Opus 5191 countVersion not specifiedon_target未注明6h timeout · ? 次尝试2026-07-24来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Opus 5171 countVersion not specifiedon_target未注明2h timeout cutoff · ? 次尝试未注明来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Opus 4.8120 countVersion not specifiedon_target未注明6h timeout · ? 次尝试2026-05-28来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
Claude Opus 4.880 countVersion not specifiedon_target未注明2h timeout cutoff · ? 次尝试未注明来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GLM-5.2 · DoGNAVY79 countVersion not specifiedon_targetDoGNAVYselected subset · ? 次尝试2026-08-22来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GLM-5.3 · Claude Code130 countVersion not specifiedon_targetClaude Coderescaled 6h timeout · ? 次尝试2026-08-18来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GLM-5.3 · Claude Code105 countVersion not specifiedon_targetClaude Coderescaled 2h timeout · ? 次尝试未注明来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GLM-5.2 · Claude Code39 countVersion not specifiedon_targetClaude Coderescaled 6h timeout · ? 次尝试2026-08-18来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明
GLM-5.2 · Claude Code29 countVersion not specifiedon_targetClaude Coderescaled 2h timeout · ? 次尝试未注明来源行日期；含义未明确 · ExploitGym website · 执行方: 未注明

范围与归属

数据报告方
ExploitGym website
实际评测方
各观测的执行方未完整确定
抓取版本
2026-10-26
来源身份
8c8a725ba1775411

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。

来源发现

以下陈述整理自原始来源，未作为 NUWA 独立实验结论。

榜单的成功标准要求利用给定的目标漏洞；仅捕获标志是另一项结果，不能等同于完成目标漏洞利用。

exploitgym.html:123-127,212-214

结论范围限于原始来源的任务与实验条件；原文定位保留供核查。

官方页面说明，评测通过面向安全研究的可信访问计划开展。

exploitgym.html:124-125

结论范围限于原始来源的任务与实验条件；原文定位保留供核查。

官方报告，启用防护措施后，成功利用的数量减少，但未降至零。

exploitgym.html:149-153

结论范围限于原始来源的任务与实验条件；原文定位保留供核查。

作者指出，自动化漏洞利用评测具有双重用途，因此需要以负责任的方式严格衡量相关能力。

exploitgym.html:157-161

结论范围限于原始来源的任务与实验条件；原文定位保留供核查。

官方正文认为，更强的智能体可能受益于更长的执行预算；相关研究条件尚未完整说明。

exploitgym.html:262-266

结论范围限于原始来源的任务与实验条件；原文定位保留供核查。

官方说明，基准设计和实验方法由学术作者开发，行业合作方提供模型访问权限和反馈。

exploitgym.html:350-353

结论范围限于原始来源的任务与实验条件；原文定位保留供核查。

提取范围

以下条目说明来源各部分的处理状态。个别观测已核验，不代表整份文档的全部结论和结果均已验证。

查看 14 项覆盖记录
章节处理状态定位与范围
Title, authors, affiliations and citationqualitativeexploitgym.html:37-77,340-353Identity and methodological attribution retained; page provides citation year 2026 but no precise page publication or revision date.
Benchmark scope and task inventoryextractedhttps://www.cybergym.io/assets/data/exploitgym.json#/instances; exploitgym.html:169-214All four exact inventory counts retained as catalogue metadata, not row denominators.
All primary submission fieldsextractedhttps://www.cybergym.io/assets/data/exploitgym.json#/results/0..1415 primary objects: 14 visible and 1 hidden. Every reported score and usage leaf, including explicit nulls, retained.
All six nested budget variantsextractedhttps://www.cybergym.io/assets/data/exploitgym.json#/results/{0,9,10,11,13,14}/subRows/0Separate contexts; parent inheritance is explicit. 2h cutoff, 4h, 6h, rescaled and selected-subset conditions stay distinct.
Visibility, rank metric, version display, averages and inheritanceextractedexploitgym.html:360-513; leaderboard.js:225-238hidden rows filtered; total prefers on_target; inherited parent metadata preserved; average presentation rounding not extracted as new values.
Key Takeawaysqualitativeexploitgym.html:133-163Only high-level conclusions retained; operational examples excluded.
Flag capture versus intended vulnerabilityextractedexploitgym.html:229-239Narrative numbers separately quarantined; GPT-5.5 conflict and hidden Mythos Preview recorded.
Different models find different exploitsextractedexploitgym.html:243-254Five explicit overlap counts retained with unresolved population; no figure estimates or inferred model totals.
More budget helpsextractedexploitgym.html:257-267Two exact Mythos narrative endpoints, approximate Opus plateau and stated time bound retained as unresolved narrative evidence.
Overview, flag, overlap and time-curve imagesfigure-onlyexploitgym.html:192,231,253,260Image files not fetched/read; no bar-height or curve estimates. Exact adjacent prose values extracted.
Example vulnerability and exploit trajectorynot-relevantexploitgym.html:274-309Operational exploit/task details and incidental example-specific numbers deliberately excluded from extraction.
Why This Mattersqualitativeexploitgym.html:311-333High-level defense and responsible-evaluation implications retained.
External system cards, reports and submissionsreusedhttps://www.cybergym.io/assets/data/exploitgym.json#/results/*/source_urlSource links preserved as dependencies; none independently read or counted as replication in this unit.
Exact executor, trials, row denominator, date semantics and missing fieldsnot-disclosedhttps://www.cybergym.io/assets/data/exploitgym.json#/resultsNo inference from model brand, reporting party, dates, absent metrics or catalogue size.

来源待核实事项

正文和结构化榜单部分统计口径未完全对应，保留各自范围，不据此合并或修正数值。

部分提交未披露版本、样本分母或执行方；重标定预算条件与普通时限条件分别保留。

版本指纹
document

https://www.cybergym.io/exploitgym/

97db69cc276fe24ba54f6032c07f0170cc889bd88368d68c65ad97ab755b0889
data

https://www.cybergym.io/assets/data/exploitgym.json

1cf3b94633364a7cb76c35dc502f21fbf889ee02b639f3a20042960a0a9b2533
renderer

https://www.cybergym.io/assets/js/leaderboard.js

7aa2d51c07d2bda8dc8be9910844441aabcfaa03b8b1a9adf1a24e3be45298f2

参考文献

数值结果由所引研究的作者报告。

  1. [1] ExploitGym official leaderboard and benchmark page