AI Image
Google announces Gemini 4 and says it’s so capable that only ‘trusted cyber defenders’ can have it right now
The Verge • • general
Key Points:
- Google has unveiled its new AI model, Gemini 4 Argon, designed to excel in complex workflows such as software engineering, enterprise knowledge work, and cybersecurity defense.
- Initial access to Gemini 4 Argon is restricted to a select group of trusted cyber defenders, with Google collaborating with the U.S. government on a voluntary pre-release access process.
- The model is already in use internally at Google, aiding in tasks like large-scale codebase migrations, and has demonstrated superior performance compared to competitors like OpenAI and Anthropic in benchmark tests.
- Google plans to implement robust safeguards against misuse, prompt injection attacks, and misalignment before a wider release of Gemini 4 Argon.
- The announcement follows closely after OpenAI’s DevDay event and coincides with Koray Kavukcuoglu’s recent appointment as head of DeepMind.
