Google rolls out Gemini 4 Argon worldwide, tops Vals Index

Google frames Argon around "deep work" — complex, multi-step missions in fields like coding, finance, legal, cybersecurity and science. The company says the model outperforms other frontier models across benchmarks and took first place on the Vals Index, which measures performance in major economic sectors including finance, coding, legal and tax. Google also points to potential uses in detecting and fixing critical cybersecurity vulnerabilities, as well as in quantum computing, nuclear fusion and life sciences.

Internally, Google says Argon is already in use: in its coding division, agent-driven code submissions rose by more than 35%, while code rollbacks fell to roughly one-third. On the infrastructure side, Google reports a 2.85x end-to-end inference speedup on A100 environments via vLLM while keeping 100% token parity and zero quality loss.

Google describes Argon as combining advanced reasoning and multimodal capabilities with strict foundational knowledge, giving it strong performance in advanced multimodal document understanding, long-term data analysis and interactive analysis sandboxes. The rollout spans AI Studio and the Gemini API for developers and enterprise users, Google Antigravity for Google AI Ultra users, and the Gemini app and AI Mode in Google Search for Ultra and Pro users.