Google has officially launched its Gemini 4 Argon AI model worldwide, offering access through the Gemini app, Search, and API to anyone willing to pay. The company introduced discounted token pricing for early adopters alongside tiered consumer subscriptions.
- Gemini 4 Argon available globally with pay-per-use API and subscription access
- Launch pricing offers significant discounts for developers and consumers
- Cybersecurity applications tested under Google’s Fairwind program
What happened
Google released Gemini 4 Argon, a next-generation AI model, accessible worldwide through the Gemini app, Google Search, and its API. Developers can pay $2 per million input tokens and $10 per million output tokens during the launch period, after which prices double. Consumers can access the model through a Pro subscription starting at $19.99 per month.
Initially, the model was released to specialized cybersecurity defenders via the Fairwind program without usual guardrails; this group has since expanded to hundreds of partners. The broader public launch enables developers and consumers to utilize the model for various advanced tasks.
Why it matters
Google describes Gemini 4 Argon as a formidable AI, ranking first on the Vals Index for performance in critical domains like finance, coding, legal, and tax tasks. The model’s cybersecurity testing revealed real-world vulnerabilities, highlighting its practical value in security and infrastructure protection.
To maintain safety and reliability, Google incorporates extensive monitoring and internal defenses, reducing success rates of prompt injection attacks to under 1%. Additionally, Google's agent tool Antigravity leverages Argon to improve coding efficiency and reduce development rollback, signaling enhanced productivity benefits.
What to watch next
Observers should monitor adoption rates of Gemini 4 Argon across industries and sectors, especially how the expanded cybersecurity partnerships use its capabilities to identify threats. The pricing model and subscription tiers, including the higher-end Ultra tier needed for advanced features, may shape user engagement and ecosystem growth.
Regulatory attention is expected as AI systems with capabilities for cyber offense, manipulation, and other systemic risks come under scrutiny. Google’s compliance with emerging AI governance frameworks and transparency on performance and safety metrics will be critical in the model’s trajectory.