{"id":96950,"date":"2026-05-08T17:53:35","date_gmt":"2026-05-08T14:53:35","guid":{"rendered":"https:\/\/forklog.com\/en\/?p=96950"},"modified":"2026-05-08T17:55:17","modified_gmt":"2026-05-08T14:55:17","slug":"openai-introduces-a-rival-to-mythos-and-three-voice-models","status":"publish","type":"post","link":"https:\/\/forklog.com\/en\/openai-introduces-a-rival-to-mythos-and-three-voice-models\/","title":{"rendered":"OpenAI Introduces a Rival to Mythos and Three Voice Models"},"content":{"rendered":"

OpenAI has unveiled<\/a> GPT-5.5-Cyber, a modification of its flagship AI model tailored for cybersecurity tasks. It is available to a limited group of developers.<\/p>\n

The neural network is not designed for significant expansion of capabilities in the digital environment; rather, it has been trained for more flexible handling of security tasks: identifying and classifying vulnerabilities, verifying fixes, and analyzing malware.<\/p>\n

\n

\u201cGPT\u20115.5\u2011Cyber will allow a limited group of partners to explore complex workflows where access specificity matters,\u201d the startup’s blog states.<\/p>\n<\/blockquote>\n

The release by OpenAI comes a month after the launch<\/a> of Mythos, an AI model from Anthropic that garnered significant attention for its cybersecurity capabilities. The firm refrained from releasing the neural network to the public due to high security risks. Instead, it introduced Project Glasswing, an initiative for testing the tool in secure conditions.<\/p>\n

In April, OpenAI released<\/a> GPT-5.4-Cyber in response, aimed at finding vulnerabilities as part of the Trusted Access for Cyber program launched in February.<\/p>\n

Voice Models<\/h2>\n

The company did not stop there. OpenAI introduced three voice models:<\/p>\n