Google Rolls Out Gemini 4 Argon Model for Multi-Prompt Tasks
On September 30, 2026, Google announced the rollout of Gemini 4 Argon, a model designed for jobs that require multiple prompts. The company claims it is state-of-the-art in various benchmarks and has already been used by thousands of employees for tasks such as coding and research.
The model will be initially available to trusted cyber defenders through the Fairwind Program, with developers, enterprises, and consumers following soon after 'as soon as possible'. Google's senior vice president Koray Kavukcuoglu stated that this approach allows them to put a strong model in the hands of defenders quickly.
The price for using Gemini 4 Argon is $2 per million input tokens and $10 per million output tokens, with cached input offered at 95% off. After the introduction period ends, these prices will increase to $4 and $20 respectively. The output token limit has been raised from 64,000 to 1 million, which Google notes is a significant increase but not necessarily a context window.
Rivals such as OpenAI have been mentioned in reports, with CNBC noting that the introductory prices for Gemini 4 Argon match those of OpenAI's discounted GPT-6.1 Sol. However, there has been no official response from Microsoft or Amazon regarding Google's new model.