Google's New Gemini AI Model Can Process Invoices, Contracts, and Taxes

Google has introduced Gemini 4 Argon, its latest AI model aimed at simplifying complex administrative tasks. The company states that Argon excels in areas such as legal research, tax preparation, and financial analysis.

According to Google, Argon has achieved a top ranking on a significant benchmark for knowledge work tasks, surpassing competitors like Anthropic's Claude and OpenAI's GPT-6. A key feature is the model's expanded output limit, increased to 1 million tokens from 64,000, allowing it to process and generate longer, more complex responses in a single interaction.

This enhanced capacity enables the model to handle extensive documents and complex reasoning without requiring tasks to be broken down. For businesses, this could translate to more efficient workflows and higher quality outputs. However, Google's introductory pricing of $2 per million input tokens and $10 per million output tokens indicates that extensive usage could incur significant costs.

The Argon model is already being utilized internally at Google for tasks like optimizing quantum computing algorithms and reducing data center memory usage. Google also highlights its enhanced cybersecurity capabilities, including the ability to identify and remediate software vulnerabilities. Broader access to Argon is currently restricted, with initial availability limited to vetted partners for cybersecurity tool usage.