Google has launched Gemini 4 Argon, its latest flagship artificial intelligence model, giving developers and businesses worldwide access to a system designed for complex, multi-step tasks.
The company says Argon is particularly capable in areas including coding, finance, law, cybersecurity and scientific research.
The launch is part of Google’s effort to strengthen its position in the increasingly competitive AI market, where it faces rivals including OpenAI and Anthropic.
Developers can access Argon through Google’s API at an introductory price of US$2 per million input tokens (about R34) and US$10 per million output tokens (about R170). The promotional pricing runs until 30 November, after which the rates will increase to US$4 and US$20 respectively.
Paying Google AI Pro and Ultra subscribers can also use Argon through the Gemini app and AI Mode in Search, while access through Google’s Antigravity coding tool is restricted to Ultra subscribers.
Google said its engineering teams using Argon through Antigravity increased AI-generated code submissions by more than 35% while reducing code rollbacks by almost a third.
Early customers testing the model include Deloitte, LexisNexis, Ambience Healthcare and Isomorphic Labs, Alphabet’s drug-discovery company.
Google is also expanding its Fairwind programme, which provides vetted cybersecurity defenders with a version of Argon without its cyber safety guardrails. The company previously made the model available to selected cybersecurity partners.
In a separate initiative, Google plans to give 10,000 scientists six months of free access to the highest usage tier of its Antigravity platform.
The launch follows a period of disruption in Google’s AI division, including delays to planned model releases and changes to the leadership structure at DeepMind.
With Argon, Google is positioning its latest model as a tool for professional and technical work as the competition to build increasingly capable AI systems intensifies.
Main Image: Red Hot Cyber
