Google Launches Gemini 4 Argon With a 1 Million-Token Output Limit
Google is starting with trusted cyber defenders before expanding access to developers, businesses, and consumers.
Gemini 4 Argon is Google’s new frontier AI model for complex work across software engineering, enterprise knowledge tasks, and cybersecurity. Google says Argon can generate up to 1 million output tokens in a single response, up from 64,000 tokens in the previous generation. Google’s official announcement details the model and its rollout.
Gemini 4 Argon Is Built for Long, Multi-Step Work
Google DeepMind says Gemini 4 Argon is designed to sustain long workflows across software engineering, legal and finance work, and cybersecurity. Google says the model can autonomously find, validate, and patch critical software vulnerabilities.
The 1 Million-Token Limit Is the Standout Change
The 1 million-token output ceiling is a major increase from the 64K output limit of the previous Gemini generation. Google reports 84.2% on its GraphWalks test for the 256K-to-1M range, compared with 99.7% up to 128K.
Argon is not launching with broad public access. Google says it is first rolling out to trusted cyber defenders through Fairwind. Broader access is planned for paid API customers and Google AI Ultra subscribers, followed by wider availability for developers, enterprises, and consumers, but Google has not given a public date.
Access Starts With Cybersecurity Defenders
Argon is not launching with broad public access. Google says it is first rolling out to trusted cyber defenders through its Fairwind program. Broader access is planned for paid API customers and Google AI Ultra subscribers, followed by wider availability for developers, enterprises, and consumers, but Google has not given a public date.
The cautious rollout follows the model’s cybersecurity capabilities. Google says Argon is designed to resist indirect prompt-injection attacks and includes monitoring that can stop execution when necessary. This continues the security focus seen in the recent Gemini 3.6 Flash and Gemini 3.5 Flash-Lite launch.
Pricing Starts at $2 Per Million Input Tokens
Google lists introductory API pricing at $2 per million input tokens and $10 per million output tokens, with cached input priced at 95% off the input rate. Saganote’s Gemini pricing guide covers the broader Google AI subscription tiers.
Google is positioning Gemini 4 Argon around tasks that require many reasoning steps, large amounts of context, and substantial output. The launch adds another major model to the growing market for agentic coding and enterprise AI, alongside Anthropic’s Claude Sonnet 5 launch.
For now, the defining change is its scale of output. Google has not announced a firm date for general Gemini 4 Argon access, so the widest release remains ahead.