{"id":78,"date":"2026-05-01T15:32:27","date_gmt":"2026-05-01T15:32:27","guid":{"rendered":"https:\/\/paltechnews.com\/index.php\/2026\/05\/01\/gpt-5-5-matches-heavily-hyped-mythos-preview-in-new-cybersecurity-tests\/"},"modified":"2026-05-01T15:32:27","modified_gmt":"2026-05-01T15:32:27","slug":"gpt-5-5-matches-heavily-hyped-mythos-preview-in-new-cybersecurity-tests","status":"publish","type":"post","link":"https:\/\/paltechnews.com\/index.php\/2026\/05\/01\/gpt-5-5-matches-heavily-hyped-mythos-preview-in-new-cybersecurity-tests\/","title":{"rendered":"GPT-5.5 matches heavily hyped Mythos Preview in new cybersecurity tests"},"content":{"rendered":"<div>\n<p>Last month, Anthropic <a href=\"https:\/\/red.anthropic.com\/2026\/mythos-preview\/\">made a big deal<\/a> about the supposedly outsize cybersecurity threat represented by its Mythos Preview model, leading the company to <a href=\"https:\/\/arstechnica.com\/ai\/2026\/04\/anthropic-limits-access-to-mythos-its-new-cybersecurity-ai-model\/\">restrict the initial release to \u201ccritical industry partners.\u201d<\/a> But <a href=\"https:\/\/www.aisi.gov.uk\/blog\/our-evaluation-of-openais-gpt-5-5-cyber-capabilities\">new research from the UK&#8217;s AI Security Institute<\/a> (AISI) suggests that OpenAI&#8217;s GPT-5.5, which <a href=\"https:\/\/openai.com\/index\/introducing-gpt-5-5\/\">launched publicly last week<\/a>, reached &#8220;a similar level of performance on our cyber evaluations&#8221; as Mythos Preview, which the group <a href=\"https:\/\/arstechnica.com\/ai\/2026\/04\/uk-govs-mythos-ai-tests-help-separate-cybersecurity-threat-from-hype\/\">evaluated last month<\/a>.<\/p>\n<p>Since 2023, the AISI has run a variety of frontier AI models through 95 different <a href=\"https:\/\/www.eccouncil.org\/cybersecurity-exchange\/ethical-hacking\/capture-the-flag-ctf-cybersecurity\/\">Capture the Flag challenges<\/a> designed to test capabilities on cybersecurity tasks, such as reverse engineering, web exploitation, and cryptography. On the highest-level &#8220;Expert&#8221; tasks, GPT-5.5 passed an average of 71.4 percent, slightly higher than the 68.6 percent achieved by Mythos Preview (though within the margin of error). In one particularly difficult task that involved building a disassembler to decode a Rust binary, AISI notes that &#8220;GPT-5.5 solved the challenge in 10 minutes and 22 seconds with no human assistance at a cost of $1.73&#8221; in API calls.<\/p>\n<p>GPT-5.5 also matched Mythos Preview in its progress on <a href=\"https:\/\/arxiv.org\/abs\/2603.11214\">&#8220;The Last Ones&#8221;<\/a> (TLO), an AISI test range set up to simulate a 32-step data extraction attack on a corporate network. GPT-5.5 succeeded in 3 of 10 attempts on TLO, compared to 2 of 10 for Mythos Preview\u2014no previous model had ever succeeded at the test even once. But GPT-5.5 still fails at AISI&#8217;s more difficult &#8220;Cooling Tower&#8221; simulation of an attempted disruption of the control software for a power plant, as every previously tested AI model also has.<\/p>\n<p><a href=\"https:\/\/arstechnica.com\/ai\/2026\/05\/amid-mythos-hyped-cybersecurity-prowess-researchers-find-gpt-5-5-is-just-as-good\/\">Read full article<\/a><\/p>\n<p><a href=\"https:\/\/arstechnica.com\/ai\/2026\/05\/amid-mythos-hyped-cybersecurity-prowess-researchers-find-gpt-5-5-is-just-as-good\/#comments\">Comments<\/a><\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>Last month, Anthropic made a big deal about the supposedly outsize cybersecurity threat represented by its Mythos Preview model, leading<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"colormag_page_container_layout":"default_layout","colormag_page_sidebar_layout":"default_layout","footnotes":""},"categories":[29,76,77,78,5,79,80,81,75,82],"tags":[],"class_list":["post-78","post","type-post","status-publish","format-standard","hentry","category-ai-gadgets","category-aisi","category-anthropic","category-cybersecurity-gadgets","category-gadgets","category-gpt-5-5","category-mythos","category-openai","category-security","category-test"],"_links":{"self":[{"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/posts\/78","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/comments?post=78"}],"version-history":[{"count":0,"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/posts\/78\/revisions"}],"wp:attachment":[{"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/media?parent=78"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/categories?post=78"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/paltechnews.com\/index.php\/wp-json\/wp\/v2\/tags?post=78"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}