{"id":411856,"date":"2026-02-06T22:05:17","date_gmt":"2026-02-06T22:05:17","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/411856\/"},"modified":"2026-02-06T22:05:17","modified_gmt":"2026-02-06T22:05:17","slug":"openais-new-model-leaps-ahead-in-coding-capabilities-but-raises-unprecedented-cybersecurity-risks","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/411856\/","title":{"rendered":"OpenAI\u2019s new model leaps ahead in coding capabilities\u2014but raises unprecedented cybersecurity risks"},"content":{"rendered":"<p>OpenAI believes it has finally pulled ahead in one of the most closely watched races in artificial intelligence: AI-powered coding. Its newest model, GPT-5.3-Codex, represents a solid advance over rival systems, showing markedly higher performance on coding benchmarks and reported results than earlier generations of both OpenAI\u2019s and Anthropic\u2019s models\u2014suggesting a long-sought edge in a category that could reshape how software is built. <\/p>\n<p>But the company is rolling out the model with unusually tight controls and delaying full developer access as it confronts a harder reality: The same capabilities that make GPT-5.3-Codex so effective at writing, testing, and reasoning about code also raise serious cybersecurity concerns. In the race to build the most powerful coding model, OpenAI has run headlong into the risks of releasing it.<\/p>\n<p>GPT-5.3-Codex is available to paid ChatGPT users, who can use the model for everyday software development tasks such as writing, debugging, and testing code through OpenAI\u2019s Codex tools and ChatGPT interface.<\/p>\n<p>But for now, the company is not opening unrestricted access for high-risk cybersecurity uses, and OpenAI is not immediately enabling full API access that would allow the model to be automated at scale. Those more sensitive applications are being gated behind additional safeguards, including a new trusted-access program for vetted security professionals, reflecting OpenAI\u2019s view that the model has crossed a new cybersecurity risk threshold.<\/p>\n<p>The company is also making $10 million in API credits available to developers who want to use OpenAI\u2019s models to work on applications that would bolster cyber defenses.<\/p>\n<p>The company\u2019s <a aria-label=\"Go to https:\/\/openai.com\/index\/introducing-gpt-5-3-codex\/\" href=\"https:\/\/openai.com\/index\/introducing-gpt-5-3-codex\/\" rel=\"nofollow noopener\" target=\"_blank\">blog post<\/a> accompanying the model release on Thursday said that while it does not have \u201cdefinitive evidence\u201d the new model can fully automate cyberattacks, \u201cwe\u2019re taking a precautionary approach and deploying our most comprehensive cybersecurity safety stack to date. Our mitigations include safety training, automated monitoring, trusted access for advanced capabilities, and enforcement pipelines including threat intelligence.\u201d<\/p>\n<p>OpenAI CEO Sam Altman posted on <a aria-label=\"Go to https:\/\/fortune.com\/company\/twitter\/\" target=\"_blank\" href=\"https:\/\/fortune.com\/company\/twitter\/\" rel=\"nofollow noopener\">X<\/a> about the concerns, saying that GPT-5.3-Codex is \u201cour first model that hits \u2018high\u2019 for cybersecurity on our preparedness framework,\u201d an internal risk classification system OpenAI uses for model releases. In other words, this is the first model OpenAI believes is good enough at coding and reasoning that it could meaningfully enable real-world cyber harm, especially if automated or used at scale.<\/p>\n<p>OpenAI\u2019s \u201cPreparedness Framework\u201d says the company will not release a model that is rated \u201chigh\u201d risk in any of the risk areas the framework addresses, including cybersecurity, without first implementing safeguards to mitigate those risks. The framework mentions a trusted access program as one potential safeguard that can be implemented.<\/p>\n<p>Feb. 6: This story has been updated to include information about the $10 million in API credits OpenAI is offering for those working on cybersecurity defenses and to include more information about OpenAI\u2019s risk assessment levels and Preparedness Framework.<\/p>\n","protected":false},"excerpt":{"rendered":"OpenAI believes it has finally pulled ahead in one of the most closely watched races in artificial intelligence:&hellip;\n","protected":false},"author":2,"featured_media":411857,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,9218,65511,5215,874,10108,86,56,54,55],"class_list":["post-411856","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-coding","tag-cyber","tag-hacking","tag-openai","tag-sam-altman","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/411856","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=411856"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/411856\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/411857"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=411856"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=411856"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=411856"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}