After months of sitting on the sidelines of the frontier artificial intelligence race, Google has launched its long-awaited Gemini 4 Argon model. The company states that the new model can compete with the industry’s best offerings.
Performance and internal applications
Google reports that Argon delivers frontier reasoning and improved performance across complex workflows while reducing hedging, repetition, and clichés. According to the company, the model outperforms competing systems such as OpenAI’s GPT-6 Astra and Anthropic’s Fable 5.1 and Opus 5.5 across benchmarks including coding, cybersecurity, science, finance, and law.
Google has integrated Argon into several internal operations. The company is using the model to find and patch cybersecurity vulnerabilities, reduce the computing power required for YouTube’s comments section by half, and debug the drone physics simulator for Alphabet’s delivery subsidiary Wing.
Availability and pricing
Argon is rolling out to higher-tier Pro and Ultra subscribers within the Gemini app and AI mode in Search, as well as to Ultra subscribers on the agentic development platform Antigravity. Availability for Plus subscribers or free users remains unconfirmed.
The model is launching with an introductory price of $2 per million input tokens and $10 per million output tokens. This rate is approximately one-fifth the cost of OpenAI’s GPT-6 Astra and Anthropic’s Fable 5 models. These introductory prices are scheduled to double at the end of November, though token costs do not necessarily translate directly to identical operational costs across different models.
Release context
Argon is Google’s first major frontier release since Gemini 3 debuted in late 2025. The public rollout follows a limited release to a select group of cyber defenders for risk testing, and arrives shortly after Meta announced its Muse Wonder Full AI model.