Google's Gemini 4 Model Identified as 'Argon'

VERA Brief

AI-generated. Grounded in the article and its cited sources.

Google's first Gemini 4 model is codenamed 'Argon' and is designed for knowledge work, cybersecurity defense, and creative writing. This development highlights Google's efforts to improve its AI capabilities.

Key facts

• Google's initial Gemini 4 large language model has the internal codename 'Argon'.
• The 'Argon' model is trained for knowledge-based tasks, cybersecurity defense, and creative writing.
• Google's ongoing efforts aim to enhance its AI offerings across multiple domains.
• The development is relevant to the verifiable accuracy of AI systems in representing and generating content.