Google Announces Gemini 4 Argon with 1M Token Output and Cybersecurity Focus

Google has announced Gemini 4 Argon, a new frontier model designed for complex, long-horizon workflows in software engineering, enterprise knowledge work, and cybersecurity defense. The model is currently rolling out to a select group of trusted cyber defenders through the Fairwind Program while Google engages in the U.S. government’s voluntary pre-release access process. Wider availability for developers, enterprises, and consumers is planned for soon, starting with paid API customers and Google AI Ultra subscribers.

Key Capabilities and Benchmarks

Gemini 4 Argon significantly expands its output token limit to an industry-leading 1 million tokens, up from the previous 64K, enabling deeper reasoning in single trajectories. The model demonstrates state-of-the-art performance across several key benchmarks:

• Software Engineering: Sets a new state of the art on DeepSWE v1.1 with a score of 77.9%.
• Enterprise Knowledge: Leads the Vals Index (measuring economic impact across finance, coding, legal, and tax) and ranks #1 on AutomationBench with a score of 51.3%.
• Video Understanding: Achieves state-of-the-art performance on LVBench with a score of 91.7%.
• Cybersecurity: Ties for first place on CWE-bench v1 with a score of 68%, demonstrating strong capabilities in remediating security vulnerabilities.

Internal Impact and Use Cases

Google reports that thousands of internal employees are already using Argon, highlighting specific productivity gains:

• Quantum Computing: Argon helped optimize quantum algorithmic subroutines, beating published baselines by 40% in minutes.
• Memory Efficiency: Agents analyzed fleet-wide telemetry to identify memory optimizations, freeing up over 300 TiB of memory with estimated total savings of 500 TiB to 1 PiB.
• Code Migration: Agents are migrating C/C++ codebases to Rust, including scaling from tens of thousands of lines in libraries like re2 and libgav1 to 800K+ lines for the Fuchsia OS Zircon kernel. In one case, Argon replaced 32K lines of SIMD code in libgav1, resulting in a memory-safe decoder running 2.7x faster than the previous Rust port.

Pricing and Safety

Gemini 4 Argon will launch at an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens priced at 95% off. After the introductory period, prices will rise to $4 per million input tokens and $20 per million output tokens.

Google emphasizes strengthened safeguards, including robust defenses against prompt injection attacks (leading on Gray Swan’s Indirect Prompt Injection benchmark) and misalignment monitoring that tracks chain-of-thought reasoning. For trusted defenders, Argon will be released without cyber guardrails to maximize its defensive capabilities, as demonstrated by its use by Wiz to uncover critical vulnerabilities in healthcare software that previous models missed.