Gemini 4 Argon: our next era of frontier intelligence
deepmind ·
Google DeepMind has announced Gemini 4 Argon, a new frontier AI model designed for complex, long-horizon professional tasks. The model features an industry-leading one million token output limit, enabling deep reasoning in software engineering, legal drafting, financial research, and cybersecurity defense. Internally, Google reports significant efficiency gains, including optimizing quantum algorithms and freeing up substantial memory resources across its data centers. Gemini 4 Argon achieves state-of-the-art performance on benchmarks such as DeepSWE v1.1, the Vals Index, and LVBench for long video understanding. In cybersecurity, the model autonomously identifies and patches vulnerabilities, with early demonstrations uncovering critical risks in healthcare software. To ensure safety, Google is implementing robust safeguards against misuse, prompt injection, and misalignment before broad public release. The model is currently rolling out to trusted cyber defenders through the Fairwind Program while undergoing rigorous testing. Pricing is set at $2 per million input tokens and $10 per million output tokens, with discounted rates for cached inputs.
deepmind