OpenAI unveils GPT-6 Astra with critical cyber capabilities
OpenAI has launched GPT-6 Astra, its first model rated at a critical cybersecurity capability level, with initial access limited to select organizations.
OpenAI has released GPT-6 Astra, described as its first model to reach a critical level in cybersecurity. Access is currently limited to a small group of organizations, with a wider rollout planned for ChatGPT Plus, Pro, Business, and Enterprise users.
In testing, Astra outperforms the previous GPT-5.6 Sol model on programming, science, and computer-use tasks, though it still trails Claude on some benchmarks.
The key feature and risk is Astra's critical cybersecurity rating: the model can identify and exploit zero-day vulnerabilities. OpenAI says these capabilities are restricted in the release version.
Source: ForkLog