Google officially announces the launch of Gemini 4 Argon, a frontier model designed to solve complex problems, providing access to developers, enterprises, scientists, and consumers. The model's launch marks another significant advance for Google in the field of artificial intelligence.
Global availability of Gemini 4 Argon
According to Google's announcement, Gemini 4 Argon is now available globally to developers and enterprises via the AI Studio and Gemini platform APIs, and Google AI Ultra subscribers can use it in Google Antigravity, while Google AI Ultra and Pro subscribers can access it in the Gemini app and AI Mode in Search.
Argon's performance across multiple domains
Argon excels in deep work across multiple domains, including coding, finance, law, cybersecurity, and science. Google states that Argon's depth and precision in multi-step tasks enable it to surpass other frontier models in benchmarks across domains, and it ranks first on the Vals Index, which measures model performance in key economic areas such as finance, coding, law, and tax.
Solving complex engineering tasks, transforming critical business and professional workflows, and accelerating real-world impact with researchers and domain experts.
Google
Argon's commercial pricing and use cases
Argon launches at a price of $2 per million input tokens (approximately HK$15.6), $10 per million output tokens (approximately HK$78), with a 95% discount on cached input tokens. Google's teams, enterprises, and research communities have already begun using Argon with significant impact, including solving refactoring and optimization of large codebases, optimizing business workflows in areas such as law, finance, and science, and finding and patching critical cybersecurity vulnerabilities.
We at Google use Argon to solve our toughest engineering problems, and in recent weeks, software engineers have been delivering high-quality code faster than ever before.
Google
Argon's technical innovations and future outlook
Argon shows significant improvements over previous Gemini models in spatial reasoning, mathematical precision, and visual coding. Google states that Argon can generate complex 3D visualizations and automatically create high-detail models in multi-agent visual evaluation loops, demonstrating its powerful technical capabilities. These innovations not only enhance the model's performance but also provide new possibilities for future application development.
Source: Google official announcement