Google DeepMind releases Gemini 4 Argon with domain-leading performance and multi-domain benchmarks

Google DeepMind today announced the global availability of Gemini 4 Argon, a new frontier model positioned for complex, real-world tasks across coding, finance, legal, cybersecurity, and science. Argon is accessible to developers and enterprises via the API in AI Studio and the Gemini platform, to Google AI Ultra subscribers in Antigravity, and to AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search.

Argon ranks first on the Vals Index, a benchmark that measures model performance across critical economic sectors including finance, coding, legal, and tax work. The company also disclosed an introductory pricing tier of $2 per million input tokens and $10 per million output tokens, with cached input tokens priced at a 95% discount.

Early adopters report measurable productivity gains. In Google’s Antigravity environment, teams increased agentic code submissions by more than 35% and reduced code rollbacks by nearly a third. Google’s open-source Agent Development Kit team used Argon to cut in-memory session prep time by 40% in long sessions and speed up cold imports by 21%. Researchers also achieved a 2.85x end-to-end inference speedup on Gemma 4 E2B using vLLM, maintaining full token parity and zero quality drop.

Beyond software engineering, Argon has been applied to accelerate quantum chemistry workflows. Using Argon in Antigravity, researchers made the xTB quantum chemistry package up to 2.3x faster at predicting molecular structures and 1.8x faster at conformer search, with changes contributed to the official xTB repository on GitHub.

Argon’s spatial reasoning and mathematical precision are highlighted in prototyping complex 3D visualizations, high-fidelity games, and interactive animations directly in the browser using slash commands and multi-agent loops in Antigravity.