On October 13, Google began general availability of its AI model "Gemini 4 Argon." It is accessible to Google AI Pro/Ultra subscribers within the Gemini app and Google Search's AI mode.
The model is designed for multi-step coding, research, and tasks in finance, legal, and scientific fields. It enhances the ability to process large volumes of documents and code, and to advance time-consuming challenges. According to Google, it secured the top position in the Vals Index, which measures model performance in business domains.
For general users, improvements include reducing verbose or unnecessarily repetitive answers, and better communicating progress during long tasks. The Gemini macOS app also introduces a coding screen from the "Google Antigravity" development environment, making it easier to transition from conversation to code creation.
In Gemini's chat interface, users can now switch from conversation to task execution. Even when users are away from their computers, tasks can be advanced using connected apps. The number of integrations has increased to over 40, enabling tasks such as ordering groceries, booking restaurants, and creating designs.
API pricing for developers is $2 per 1 million input tokens and $10 per 1 million output tokens, with cached input receiving a 95% discount. This introductory period lasts until November 30, after which prices will be $4 for input and $20 for output.
At its announcement, Gemini 4 Argon garnered attention by outperforming top models from OpenAI and Anthropic in some benchmarks. However, good test scores don't always translate to real-world performance. The key question is whether the capabilities demonstrated in benchmarks will be realized in practical use.
