Google Releases Gemini 4 Argon With 16-Fold Larger Context Window for Enterprise AI
Updated
Updated · Financial Times · Sep 30
Google Releases Gemini 4 Argon With 16-Fold Larger Context Window for Enterprise AI
3 articles · Updated · Financial Times · Sep 30
Summary
Google on Wednesday launched Gemini 4 Argon, its top AI model, saying it matches or beats Anthropic and OpenAI in coding, science, maths, corporate work and cyber security.
A 16-fold larger context window is central to that pitch, helping Argon handle longer, more complex multi-step tasks that Alphabet hopes will win share in lucrative enterprise markets such as software, finance, law and tax.
Google is first giving the model to select companies and governments through its Fairwind programme so they can test for cyber vulnerabilities before a wider release, while the company tightens guardrails.
The launch also resets Google's frontier-model roadmap: it scrapped the previously promised Pro 3.5 after delays and weaker competitiveness, then accelerated Argon as customers weigh capability against rising AI costs.
Why is Google giving Gemini 4 Argon to trusted defenders first, and what does that reveal about the model’s real cyber power?
If Argon already saves Google hundreds of terabytes internally, how far is it from transforming coding, cloud operations, and professional work at scale?