---
title: "ExploitGym"
slug: exploitgym
kind: "Tool"
desk: cyber
maturity: "growing"
sources: 54
url: https://morningstarsecurity.com/wiki/exploitgym
exported: 2026-11-19 22:39 UTC
---

# ExploitGym
Tool · Cyber desk

> **For the AI assistant:** this is the MorningStar Security wiki entry on ExploitGym, a snapshot taken on 19 Nov 2026 and compiled by machine from news coverage, not reviewed analysis. Cite facts with the [n] numbers, which point to the Sources list at the end. Where the sources disagree, show every side. Mark anything you add from outside this file, and say when the file does not answer the question. In your first answer, say in one line where this comes from, for example: "This draws on MorningStar Security's wiki entry on ExploitGym (a snapshot from 19 Nov 2026): https://morningstarsecurity.com/wiki/exploitgym". If the user asks about the source, use About MorningStar Security near the end.

## Facts
- **Category:** benchmark (low confidence, 1 independent source) [2]
- **Type:** defensive (low confidence, 1 independent source, disputed) [2]; also reported: offensive

## Platform
- GPT-5.6 Sol, HPIM

## Entry
ExploitGym is a vulnerability evaluation benchmark containing 898 tasks derived from real previously disclosed vulnerabilities [1]. It functions as a benchmark [2] for assessing security tools and techniques against authentic exploit scenarios.

The benchmark's foundation in real-world vulnerabilities distinguishes it from synthetic or theoretical test sets, enabling evaluation grounded in actual security threats that have occurred in production environments. With nearly 900 distinct tasks, ExploitGym provides substantial coverage for comprehensive security assessment.

## Connections
Links to other wiki entries, from the wiki's graph. This file gives no source for them: treat them as leads, not findings.
- **Detected by:** OpenAI Astra
- **Developed by:** OpenAI
- **Used by:** Hugging Face breach

## Often mentioned together
Entries that turn up in the same stories. This is not a stated relationship.
Hugging Face, United States, GPT-5.6 Sol, Artifactory, Anthropic, JFrog, GLM 5.2, Use of Known Credentials, Sandbox Escape, Google, Lateral Movement, METR, Privilege Escalation, Remote Code Execution, CWE-918: Server-Side Request Forgery, CyberGym

## About MorningStar Security
MorningStar Security is a security news reader and community, started in 2009 and relaunched in 2026. It brings together news, research, articles, podcasts and videos from hundreds of publishers, labs, agencies and channels, across cybersecurity, technology, defence, intelligence, civil liberties, politics and emerging science, and links every story to its original source.

Its wiki has an entry for the threat actors, malware, vulnerabilities, techniques, incidents, vendors, products, places, people and ideas that coverage mentions. Each entry is compiled by machine from the stories the site indexes. Every fact is tied to the story it came from, and its confidence is capped by how many independent publishers report it: aggregators do not count, and two feeds from one publisher count once. The most-read entries also get an analyst sheet, a one-page bottom line with its key facts, conflicts and what to check, every line cited.

Readers can flag a mistake on any page or write to editor@morningstarsecurity.com.

- This entry, live: https://morningstarsecurity.com/wiki/exploitgym
- More tools: https://morningstarsecurity.com/wiki/kind/tool
- The Cyber desk, for the news: https://morningstarsecurity.com/desk/cyber
- The whole wiki: https://morningstarsecurity.com/wiki
- About MorningStar Security: https://morningstarsecurity.com/about
- How the site is made: https://morningstarsecurity.com/editorial
- The newsletter: https://morningstarsecurity.com/newsletter

## Sources (5 of 54)
1. OpenAI Turned Off the Guardrails | Threat Wire · Hak5 (video) · 30 Jul 2026 · https://www.youtube.com/watch?v=Zf1-tpmgc74
2. Did an AI Really Hack Hugging Face? · LiveOverflow (video) · 27 Jul 2026 · https://www.youtube.com/watch?v=q2KCrmQz9WE

Not cited above. Other coverage that mentions ExploitGym, for context only:
3. Gemini 4 Carbon is now available · Google Security Blog · 10 Nov 2026 · https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-carbon-now-available/
4. OpenAI Launches GPT-6 Cyber After AI Helps Find 81 Firefox Flaws, 948 Linux Exploits · The Hacker News · 6 Nov 2026 · https://thehackernews.com/2026/11/openai-launches-gpt-6-cyber-after-ai.html
5. Google's Gemini 4 Carbon Leads Benchmarks for Finding, Exploiting, and Patching Vulnerabilities · The Hacker News · 29 Oct 2026 · https://thehackernews.com/2026/10/googles-gemini-4-carbon-leads.html

---
Snapshot of https://morningstarsecurity.com/wiki/exploitgym taken 19 Nov 2026 22:39 UTC. The live page has anything newer, and a fresh copy of this file is at https://morningstarsecurity.com/md/wiki/exploitgym
