US EditionOpenAI Unveils GPT-5.4-Cyber a Week After Rival's Announcement of AI Model
The limited rollout gives vetted defenders fewer restrictions for binary reverse engineering and other advanced workflows as OpenAI prepares for more capable models.
- On Tuesday, OpenAI announced GPT-5.4-Cyber, a new model variant specifically designed for digital defenders. The model enables advanced defensive workflows, including binary reverse engineering, without requiring source code access.
- Access to the model is limited to the highest tier of users willing to authenticate themselves as cybersecurity defenders. OpenAI uses 'know your customer' validation systems to allow controlled access while remaining as "democratized" as possible.
- This rollout is part of an expanded Trusted Access for Cyber initiative, utilizing an "iterative deployment" process. OpenAI is "carefully" releasing capabilities to ensure "resilience to jailbreaks and other adversarial attacks, and improving defensive capabilities."
- The news follows an announcement last week by competitor Anthropic, which privately released Claude Mythos Preview. Anthropic claims the model could be exploited by bad actors if released more broadly.
- OpenAI says the initiative fits into broader security efforts, including the "Preparedness Framework" designed to defend against "severe harm from frontier AI capabilities." The company expects these safeguards to support upcoming more powerful models coming this year.
31 Articles
31 Articles
OpenAI unveils a rival to Anthropic’s Mythos
OpenAI unveiled a “Cyber” model to rival Anthropic’s Mythos. The new Anthropic model sparked consternation, especially in the banking sector, with its apparent ability to find and exploit vulnerabilities in almost any website or browser; it was only released to key organizations to let them patch any such vulnerabilities. OpenAI’s GPT-5.4-Cyber is similarly being only partially released, reinforcing an Anthropic co-founder’s warning at Semafor W…
OpenAI follows Anthropic's lead in limited release of GPT‑5.4‑Cyber
OpenAI has unveiled GPT-5.4-Cyber, a new AI model that may be willing to accept seemingly malicious prompts in the name of cybersecurity. Fortunately, the ChatGPT developer won't let just anyone play with its less restrictive, more freewheeling AI. SEE ALSO: Is Anthropic's Claude Mythos a big stunt, or a real security threat? What the experts say. Announced via a blog post on Tuesday, GPT-5.4-Cyber is a variant of OpenAI's publicly available GP…
OpenAI, the developer of ChatGPT, has begun providing a new artificial intelligence (AI) model specialized in detecting security vulnerabilities to select companies. This move comes just a week after competitor Antropic unveiled 'Mythos,' known for its superior security performance, and OpenAI has introduced a similar security-dedicated model in response.
OpenAI GPT-5.4-Cyber Model Launched to Enhance Defensive Security Work Following Anthropic’s Mythos Launch | 📲 LatestLY
OpenAI has launched GPT-5.4-Cyber, a fine-tuned model for defensive security, following the debut of Anthropic’s Mythos AI. Available through a tiered "Trusted Access for Cyber" program, the model allows vetted professionals to perform advanced tasks like reverse engineering with fewer restrictions. 📲 OpenAI GPT-5.4-Cyber Model Launched to Enhance Defensive Security Work Following Anthropic’s Mythos Launch.
Coverage Details
Bias Distribution
- 46% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
