ExploitGym

An exploit-development benchmark built from real vulnerabilities in userspace programs, V8, and the Linux kernel, requiring code execution through the intended vulnerability

Security and operationsIntended-exploit success rate

Stephen M. Walker II · Co-Founder / CEO · November 5, 2026 · 1 min read

What does ExploitGym measure?

ExploitGym tests whether an agent can develop a working exploit for a specified, real-world vulnerability. Its targets include userspace programs, Google's V8 engine, and the Linux kernel. Success requires obtaining the target's flag through code execution and verifying that the exploit used the intended vulnerability.

The benchmark's current public release contains 869 instances. The provider reports below do not all identify their task revision or subset, so that public-release count does not establish the denominator of an individual result.

OpenAI release results

OpenAI's November 5, 2026 launch report gives GPT-6 Cyber a 48.5% success rate, compared with 32.1% for GPT-5.6 Cyber. These are the launch report's rounded values. Both models have restricted cybersecurity access; the availability links identify their access terms.

Model configurationScoreHarness and setupSource and date
GPT-6 Cyber (Effort unspecified)Approved Daybreak Red organizations48.5%Harness and effort not specifiedOpenAI, 2026-11-05; Provider-reported launch values; restricted cyber models
GPT-5.6 Cyber (Effort unspecified)Approved Daybreak Red organizations32.1%Harness and effort not specifiedOpenAI, 2026-11-05; Provider-reported launch values; restricted cyber models
OpenAIExploitGym; target vulnerabilities providedIntended-exploit success rateopenai.com

OpenAI system-card results

The GPT-6 Cyber system card, section 7.0.5, reports 48.45% for GPT-6 Cyber and 42.43% for GPT-6 Astra. These values retain the card's precision and comparison cohort. The card does not identify each row's reasoning effort or run harness.

Model configurationScoreHarness and setupSource and date
GPT-6 Cyber (Effort unspecified)Approved Daybreak Red organizations48.45%Harness and effort not specifiedOpenAI, 2026-11-05; System card section 7.0.5; precision retained as published
GPT-6 Astra (Effort unspecified)42.43%Harness and effort not specifiedOpenAI, 2026-11-05; System card section 7.0.5; precision retained as published
OpenAIExploitGym, standard mitigationsIntended-exploit success ratedeploymentsafety.openai.com

Hardened evaluation

The same card's section 7.0.6 reports a separate hardened variant. It adds stronger userspace exploit mitigations, enables the V8 sandbox, and uses stricter kernel configurations where supported. These scores measure a different setting from standard ExploitGym.

Model configurationScoreHarness and setupSource and date
GPT-6 Cyber (Effort unspecified)Approved Daybreak Red organizations36.94%Stronger userspace, V8 sandbox, and kernel mitigationsOpenAI, 2026-11-05; System card section 7.0.6; distinct task setting
GPT-6 Astra (Effort unspecified)29.69%Stronger userspace, V8 sandbox, and kernel mitigationsOpenAI, 2026-11-05; System card section 7.0.6; distinct task setting
OpenAIExploitGym HardenedIntended-exploit success ratedeploymentsafety.openai.com

Google Carbon release result

Google's Carbon evaluation methodology specifies Antigravity, one attempt per task, no internet access, and benchmark mitigations disabled. Its October 28 launch result reflects the Fairwind evaluation configuration with guardrails disabled. The methodology discusses both six-hour and ten-hour timeouts without identifying which produced the displayed table value.

Model configurationScoreHarness and setupSource and date
Gemini 4 Carbon (High)54.8%Guardrails disabled; Fairwind evaluation configurationGoogle, 2026-10-28; Google methodology; source discusses 6-hour and 10-hour timeouts without tying the table value to one
GoogleAntigravity; one attempt per task; no internetExploitGym, benchmark mitigations offSuccess rateblog.google

How to compare these results

Compare models within a source's published evaluation setup. Differences in task revisions, time limits, harnesses, mitigations, and access configurations can change what a score measures. ExploitGym Hardened and the scope-following ExploitGym Honeypot evaluation are separate settings. ExploitBench, SEC-bench Pro, and SRE-Bench test different tasks.