Google Unveils Latest Frontier AI, Gemini 4 Argon, Igniting AI Price Competition
- Input
- 2026-10-01 10:39:19
- Updated
- 2026-10-01 10:39:19

[Financial News] Google officially announced on the 30th of last month (local time) its next-generation frontier AI model, 'Gemini 4 Argon,' which is specialized in high-performance tasks such as software engineering, knowledge work and cybersecurity.
According to Yahoo Finance, Google said it would first provide Gemini 4 Argon to member companies participating in Fairwind, a security program. This will allow external companies to identify and fix flaws in their software before the model is officially released. Google also added that it is cooperating with the U.S. government's voluntary pre-release evaluation program for frontier models.
Following the announcement, shares of Alphabet, Google's holding company, remained largely unchanged.
Google not only emphasized the model's performance but also intensified the AI price war by offering exceptionally low token prices. It set the initial price for Gemini 4 Argon at $2 per 1 million input tokens and $10 per 1 million output tokens. A token is a unit representing a word or word fragment used in prompts (inputs) and responses (outputs). This is significantly cheaper than rival offerings. Anthropic's top-tier model, Fable 5.1, currently costs $10 per 1 million input tokens and $50 per 1 million output tokens, while Opus 5.5 costs $4 and $20, respectively.
OpenAI's highest-performing model, GPT-6 Astra, also costs $10 per 1 million input tokens and $50 per 1 million output tokens. Its latest model, GPT-6.1 Sol, is priced at the same level as Gemini 4 Argon: $2 for input and $10 for output.
The launch is being interpreted as an aggressive move by Google to regain the lead in the cutting-edge AI market after it was recently viewed as having fallen behind OpenAI's GPT-6 Astra and Anthropic's Fable 5. Google explained that it is already using Gemini 4 Argon in various internal workflows, including coding, research and document creation.
However, concerns reportedly emerged within Google before the official unveiling that Gemini 4 Argon might not be as powerful as the latest products from competitors Anthropic and OpenAI. Internal counterarguments that the model is sufficiently competitive in terms of performance are also said to be holding their ground.
Google plans to roll out the model in stages before expanding it to businesses and consumers at large. In the future, Gemini 4 Argon is expected to become a foundational model for Google's core platforms, just as the previous Gemini 3 model was applied across Google's services, from Search to YouTube.
[email protected] Yoon Jae-jun Reporter