OpenAI has expanded its Daybreak cyber defence service and launched a new artificial intelligence model designed to perform specialised cybersecurity tasks, in a move that follows the recent introduction of Anthropic’s specialised Mythos model.
Daybreak provides an integrated suite of AI models, tools and workflows designed to support cybersecurity defenders. The service will be available through two tiers, Blue and Red, offering a range of capabilities including incident response, malware analysis, vulnerability detection and security testing.
OpenAI said the new model is based on GPT-5.6 Sol, with enhanced capabilities tailored to specialised cybersecurity tasks. The model is currently available exclusively to a select group of “trusted customer partners”.
The launch comes amid growing concerns over the potential use of artificial intelligence to carry out faster and more extensive cyberattacks, as organisations increasingly adopt the technology to strengthen their defensive capabilities.
OpenAI warned that the continued development of AI-enabled cyberattacks could reduce the time available for defenders to prepare for and respond to cybersecurity threats.
