Google reveals Gemini 4 Argon

Google announced Gemini 4 Argon, a new frontier AI model. Chief AI architect and Google DeepMind SVP Koray Kavukcuoglu says it delivers frontier performance in complex workflows across software engineering, enterprise knowledge work, and cybersecurity defense.

“frontier performance in complex workflows across real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense”

Access limited to trusted defenders

Google is restricting early access to Gemini 4 Argon to a set of trusted cyber defenders. Kavukcuoglu says Google is actively engaged in the U.S. government's voluntary process for pre-release model access while gradually expanding access.

“actively engaged in the U.S. government's voluntary process for pre-release model access while we gradually expand access”

Model already powers Google internally

Gemini 4 Argon is already powering Google's internal workflows. The model has helped with tasks like large-scale codebase migrations, demonstrating its capability in real-world software engineering work.

Outperforms OpenAI and Anthropic models

Google shared benchmarks showing Gemini 4 Argon fares better than competing models from OpenAI and Anthropic. The AI community had been buzzing earlier this week on X over what appeared to be leaked benchmarks for the new model.

Safety checks before broader rollout

Google will strengthen critical frontier safeguards before a broader rollout, including defending against misuse and prompt injection attacks and monitoring for misalignment. The company is making sure the model is not misaligned before wider release.

Timing follows OpenAI's DevDay

The launch comes a day after OpenAI's DevDay conference, where OpenAI launched its Dots AI agent and GPT-6.1 Sol AI model. It also arrives soon after Kavukcuoglu's appointment as the new boss of DeepMind in August.