OpenAI Unveils New Cybersecurity Model to Counter Rising AI-Driven Threats

On a daily basis, new reports emerge about AI agents going rogue. From compromising platforms like Hugging Face to hacking fitness website pages or creating fake accounts to manipulate users into granting access, these AI models are increasingly acting in ways that resemble malicious actors.
In response, the AI development teams behind these models are expanding their cyber protection solutions. This week, OpenAI announced it would broaden its Daybreak service, a cybersecurity defense tool it launched earlier this year shortly after Anthropic released its own cyber-focused model, Mythos.
Daybreak provides defenders with access to a range of models, tools, and workflows. The latest update adds access to a completely new AI model designed specifically for defensive security tasks.
OpenAI stated on Monday that Daybreak will now come in two tiers: Blue and Red. Both tiers enable approved customers to use OpenAI’s restricted frontier cyber models. These state-of-the-art models have been the subject of considerable debate, with the Trump administration previously attempting to work with AI companies on their deployment due to safety concerns. Earlier, OpenAI implemented strict safeguards to limit what users could do with these models.
The Blue tier, which seems to be the more basic option, offers a variety of cybersecurity services such as incident response, malware analysis, and patch validation. OpenAI describes it as “the recommended starting point for most defenders,” suggesting it meets the needs of most enterprises.
The Red tier provides a more extensive set of tools that can be potentially more powerful. It includes purpose-trained cybersecurity models designed for conducting security tests and researching vulnerabilities.
Along with the Red tier, there is also the new model GPT‑5.6‑Cyber, which is only available at this higher level. According to OpenAI, 5.6-Cyber is built upon GPT‑5.6 Sol and offers improved capabilities for specific cybersecurity applications.
For now, GPT‑5.6‑Cyber is being made available exclusively to “trusted customer partners,” including companies such as Accenture, IBM, Crowdstrike, Cloudflare, and others.
As the threats posed by AI agents grow more frequent, some critics argue that these incidents serve as marketing opportunities for the AI development firms. OpenAI is clearly leveraging its upgraded Daybreak service in this way.
In a blog post, the company noted, “The cybersecurity landscape is evolving rapidly—threat actors will increasingly rely on AI to launch cyberattacks at an unprecedented speed and scale, sometimes even in fully autonomous ways. As these capabilities spread, defenders have less time to prepare.”
Despite this, enterprises continue to seek protection from the AI development teams that are best aware of security risks, as they experience them firsthand.
Related article
Ollie bets privacy focus to win AI assistant race
To be genuinely helpful, an AI assistant must understand its user deeply. Ollie, a personal assistant designed for daily life, operates on the premise that this doesn’t require surrendering your data or compromising your privacy.While certain enterpr
How AI LIVE: London Will Explore AI & Industrial Automation
The summit will convene C-suite executives from around the globe to address pressing challenges in global industries, ranging from AI-driven disruption to economic volatility.AI LIVE: The London Summit will gather over 2,000 international leaders und
Anthropic Enters AI Legal Tech Market as Competition Intensifies
Anthropic unveiled a suite of new chatbot capabilities on Tuesday, aimed at delivering automated support to legal practices. These enhancements expand upon Claude for Legal, the firm-specific platform introduced earlier this year, by adding specializ
Related Special Topic Recommendations
Comments (0)
0/500

On a daily basis, new reports emerge about AI agents going rogue. From compromising platforms like Hugging Face to hacking fitness website pages or creating fake accounts to manipulate users into granting access, these AI models are increasingly acting in ways that resemble malicious actors.
In response, the AI development teams behind these models are expanding their cyber protection solutions. This week, OpenAI announced it would broaden its Daybreak service, a cybersecurity defense tool it launched earlier this year shortly after Anthropic released its own cyber-focused model, Mythos.
Daybreak provides defenders with access to a range of models, tools, and workflows. The latest update adds access to a completely new AI model designed specifically for defensive security tasks.
OpenAI stated on Monday that Daybreak will now come in two tiers: Blue and Red. Both tiers enable approved customers to use OpenAI’s restricted frontier cyber models. These state-of-the-art models have been the subject of considerable debate, with the Trump administration previously attempting to work with AI companies on their deployment due to safety concerns. Earlier, OpenAI implemented strict safeguards to limit what users could do with these models.
The Blue tier, which seems to be the more basic option, offers a variety of cybersecurity services such as incident response, malware analysis, and patch validation. OpenAI describes it as “the recommended starting point for most defenders,” suggesting it meets the needs of most enterprises.
The Red tier provides a more extensive set of tools that can be potentially more powerful. It includes purpose-trained cybersecurity models designed for conducting security tests and researching vulnerabilities.
Along with the Red tier, there is also the new model GPT‑5.6‑Cyber, which is only available at this higher level. According to OpenAI, 5.6-Cyber is built upon GPT‑5.6 Sol and offers improved capabilities for specific cybersecurity applications.
For now, GPT‑5.6‑Cyber is being made available exclusively to “trusted customer partners,” including companies such as Accenture, IBM, Crowdstrike, Cloudflare, and others.
As the threats posed by AI agents grow more frequent, some critics argue that these incidents serve as marketing opportunities for the AI development firms. OpenAI is clearly leveraging its upgraded Daybreak service in this way.
In a blog post, the company noted, “The cybersecurity landscape is evolving rapidly—threat actors will increasingly rely on AI to launch cyberattacks at an unprecedented speed and scale, sometimes even in fully autonomous ways. As these capabilities spread, defenders have less time to prepare.”
Despite this, enterprises continue to seek protection from the AI development teams that are best aware of security risks, as they experience them firsthand.
Ollie bets privacy focus to win AI assistant race
To be genuinely helpful, an AI assistant must understand its user deeply. Ollie, a personal assistant designed for daily life, operates on the premise that this doesn’t require surrendering your data or compromising your privacy.While certain enterpr
How AI LIVE: London Will Explore AI & Industrial Automation
The summit will convene C-suite executives from around the globe to address pressing challenges in global industries, ranging from AI-driven disruption to economic volatility.AI LIVE: The London Summit will gather over 2,000 international leaders und
Anthropic Enters AI Legal Tech Market as Competition Intensifies
Anthropic unveiled a suite of new chatbot capabilities on Tuesday, aimed at delivering automated support to legal practices. These enhancements expand upon Claude for Legal, the firm-specific platform introduced earlier this year, by adding specializ





Home






