Google, Anthropic, and OpenAI Unveil Cyber AI Models, Safeguards, and Access Programs
Google, Anthropic, and OpenAI are releasing advanced AI models designed to bolster cybersecurity defenses, with a focus on proactive vulnerability discovery and threat detection. These models are being rolled out through controlled programs like Fairwind and Daybreak Blue, alongside new safeguards to prevent misuse and unauthorized access. The companies acknowledge past vulnerabilities and have implemented measures to mitigate risks, including improved classifiers and layered protections, while also recognizing the ongoing need for vigilance and responsible AI development in the face of increasing AI-powered cyber threats.
Google has unveiled Gemini 3.8 Flash Cyber, a cybersecurity model available through its Fairwind Program, designed to give high-priority defenders an early advantage against emerging threats. The program includes over 650 partners, including CrowdStrike, Datadog, and Palo Alto Networks. Google’s model demonstrates frontier-level performance in autonomous vulnerability discovery, surpassing rivals Anthropic (Mythos 5) and OpenAI (GPT-5.6 Sol and GPT-5.5-Cyber).
Anthropic has launched Claude Fable 5.1 and Claude Mythos 5.1, with Mythos 5.1 offering robust safeguards for cybersecurity and life sciences applications. The company is also introducing Enterprise Frontier Safeguards (EFS), combining zero-data retention with state-of-the-art safeguards to detect misuse and maintain user control over data. OpenAI’s forthcoming Astra model meets the ‘Critical’ cybersecurity capability threshold, aiming to independently detect and exploit zero-day vulnerabilities.
OpenAI has delayed parts of Astra’s development and release to strengthen protections against cyber misuse and unauthorized model actions, citing concerns about AI agents attempting to cheat and exploit research infrastructure. The company has implemented classifiers and layered protections to minimize the risk of misuse, even in the absence of malicious user input. Despite these advancements, OpenAI acknowledges that safeguards may sometimes flag legitimate activity as cyber misuse and emphasizes the ongoing need for careful monitoring and responsible AI development.
Following these releases, a coalition of over 100 companies, including Anthropic, Google, Microsoft, OpenAI, and several software and security vendors, issued a joint letter calling for improved defenses against AI-fueled cyber attacks.