According to the provided report, Gemini 4 Carbon is Google's most advanced AI model to date. It reportedly outperformed Google's Gemini 4 Argon and OpenAI's GPT-6 Astra in benchmark tests.
Google says Gemini 4 Carbon has been designed specifically for cybersecurity experts. The model is intended to help identify vulnerabilities in computer systems and assist in fixing them.
The company is initially limiting access to the model, making it available only to selected cybersecurity professionals rather than the general public.
What Is Google's Genie 4 AI Model?
Alongside Gemini 4 Carbon, Google is introducing Genie 4, a model designed to simulate real and imagined environments.
Known as a world model, Genie 4 can create digital environments rather than simply predicting the next words in a conversation, as conventional chatbots do.
World models can be useful in several fields, including robotics and video game development, where simulated environments can support the creation and testing of digital experiences.
Google has also incorporated audio capabilities into Genie 4. Developers will be able to add background sounds and dialogue to AI-generated worlds, expanding the possibilities for creating interactive digital environments.
Are Gemini 4 Carbon and Genie 4 Available to Everyone?
Neither of the two models is currently available to general users, according to the provided report.
Access to Gemini 4 Carbon is restricted to selected cybersecurity experts. The report does not specify the availability arrangements for Genie 4 beyond stating that it is not yet accessible to ordinary users.
The introduction of the two models highlights Google's work on AI systems designed for specialised technical tasks and the generation of interactive digital worlds.
