{"id":2692,"date":"2026-09-04T11:55:24","date_gmt":"2026-09-04T11:55:24","guid":{"rendered":"https:\/\/thedigitalfortress.us\/?p=2692"},"modified":"2026-09-04T11:55:24","modified_gmt":"2026-09-04T11:55:24","slug":"gpt-6-astra-scores-100-on-exploitbench-as-openai-blocks-poc-exploit-requests","status":"publish","type":"post","link":"https:\/\/thedigitalfortress.us\/?p=2692","title":{"rendered":"GPT-6 Astra Scores 100% on ExploitBench as OpenAI Blocks PoC Exploit Requests"},"content":{"rendered":"
\n
<\/a><\/div>\n

OpenAI on Thursday officially unveiled GPT\u20116 Astra<\/a><\/strong>, which it described as the \u00abworld’s most intelligent and aligned model.\u00bb<\/p>\n

The development comes days after the artificial intelligence (AI) company said the model had reached the \u00abCritical\u00bb cybersecurity capability threshold under its Preparedness Framework.<\/p>\n

\u00abAstra is state-of-the-art on computer use, browsing, software engineering, cybersecurity, science, and professional work. Astra saturates FrontierMath Tier 4 with a 98% score,\u00bb OpenAI said. \u00abAstra also saturates ARC-AGI-3 with a 99.9% score and ExploitBench with a 100% score. It also sets a new frontier on computer and browser use, handling the most demanding professional work with unmatched speed, accuracy, and judgment.\u00bb<\/p>\n

The model is currently rolling out to a small set of organizations and is expected to be available to all ChatGPT Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API, Microsoft Azure, and Amazon Web Services (AWS) Bedrock.<\/p>\n

On ExploitBench, which evaluates a model’s ability to turn known software vulnerabilities into working exploits, Astra achieved a perfect score of 100%, as opposed to 78.5% for GPT\u20115.6 Sol, its previous frontier cyber-capable model.<\/p>\n

\n
\"Cybersecurity\"<\/a><\/div>\n<\/div>\n

OpenAI said the model also achieves substantially higher arbitrary code-execution rates than GPT\u20115.6 Sol when testing its exploit development capabilities using flaws from the previous three months between July and August 2026. This included two zero-day vulnerabilities in unspecified software.<\/p>\n

Astra is also equipped to use previously unknown vulnerabilities to achieve code execution in hardened browsers and develop privilege-escalation exploits for hardened operating-systems, if allowed to run without any safeguards.<\/p>\n

Given the dual-use nature of these tools \u2013 the capabilities that can help defenders find weaknesses faster can also be abused by bad actors to exploit them more easily \u2013 OpenAI said the version of Astra being released is limited to secure code review and patching, while refusing to comply with prompts related to creating proof-of-concept (PoC) exploits for vulnerabilities.<\/p>\n

\u00abThrough OpenAI Daybreak\u2060, we plan to expand access and roll out less restrictive safeguards in the coming weeks,\u00bb the AI upstart said. \u00abThis will enable more defensive workflows, including vulnerability and proof-of-concept validation, malware analysis, and detection engineering.\u00bb<\/p>\n

To address concerns about model misuse, OpenAI said it has included stronger model robustness to better tackle jailbreaks, more context to its monitoring systems, and extra safeguards to help detect and contain misalignment.<\/p>\n

Astra is also \u00abmore likely\u00bb to operate within the confines set by the user and implied by its environment, although it warned the safety checks can sometimes interrupt legitimate work, including defensive cybersecurity, at which point, the user will be prompted to review the action before continuing.<\/p>\n

\u00abIn sensitive environments, Astra proceeds with care commensurate with its risk,\u00bb the company added. \u00abIn an evaluation of computer use tasks adversarially selected to elicit misbehavior, Astra was more successful at avoiding unintended consequences. Running with additional security measures offered by default yielded even stronger performance.\u00bb<\/p>\n

\n
\"Cybersecurity\"<\/a><\/div>\n<\/div>\n

The release of Astra comes as OpenAI launched a new initiative designed to provide subsidized access to its models, hands-on training, and technical assistance to critical infrastructure sectors, including water systems, electricity providers, state and local governments, banks, non-profits, open-source maintainers, and organizations with limited security resources.<\/p>\n

The global project<\/a>, called Daybreak for Frontline Defenders<\/a>, aims to commit $1 billion to help defenders use frontier AI cyber capabilities to safeguard essential services against cyber attacks. In tandem, the company has announced a new pilot with the U.S. Multi-State Information Sharing and Analysis Center (MS-ISAC) to equip an initial group of public sector and water system defenders with Daybreak access, guided training, and hands-on assistance.<\/p>\n

\u00abWe have a defender’s window: a narrowing opportunity to use AI to close security gaps before attackers seize them,\u00bb OpenAI said. \u00abOur role is to help put powerful tools in defenders\u2019 hands so they can protect the systems, and the people, they are responsible for.\u00bb<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

OpenAI on Thursday officially unveiled GPT\u20116 Astra, which it described as the \u00abworld’s most intelligent and aligned model.\u00bb The development comes days after the artificial intelligence (AI) company said the…<\/p>\n","protected":false},"author":1,"featured_media":2693,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[2967,685,120,3303,3301,512,1730,1735,3302],"class_list":["post-2692","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized","tag-astra","tag-blocks","tag-exploit","tag-exploitbench","tag-gpt6","tag-openai","tag-poc","tag-requests","tag-scores"],"_links":{"self":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/posts\/2692","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=2692"}],"version-history":[{"count":0,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/posts\/2692\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/media\/2693"}],"wp:attachment":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2692"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2692"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2692"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}