Gemini 4 Argon

Google's new frontier model for long-horizon coding, enterprise knowledge work, and autonomous cyber defense, rolled out first to trusted defenders.

Google · Frontier/foundation models · model

Published in Feed

Overview

Google DeepMind announced Gemini 4 Argon on September 30, 2026 as a frontier model built for long-horizon software engineering, enterprise knowledge work, legal/finance analysis, and cybersecurity defense. Argon ties for first place on CWE-bench v1 with a 68% score for automated vulnerability remediation and was trained to autonomously find, validate, and patch critical software vulnerabilities across codebases spanning 20 programming languages. The model supports a 1,000,000-token output budget and leads Gray Swan's Indirect Prompt Injection (IPI) Benchmark for prompt-injection robustness. Google initially limited Argon, without cyber guardrails, to trusted testers and cyber defenders through its Fairwind Program and U.S. government pre-release programs. On October 13, 2026 it made Argon broadly available through the Gemini API and AI Studio, the Gemini app and AI Mode in Search for Google AI Ultra and Pro subscribers, and Antigravity, with introductory API pricing of $2 per million input and $10 per million output tokens.

This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.

Key features

  • 68% on CWE-bench v1, tying for first place on automated vulnerability remediation
  • 1,000,000-token output window for long-horizon coding and analysis tasks
  • Autonomous vulnerability discovery, validation, and patching across 20 programming languages
  • Leading score on Gray Swan's Indirect Prompt Injection (IPI) Benchmark
  • Internal activation-monitoring techniques to spot misuse
  • Staged rollout: Fairwind Program trusted defenders and U.S. government pre-release first, then paid API and Google AI Ultra
  • Broad availability from October 13, 2026 via Gemini API, AI Studio, Gemini app and AI Mode for Ultra and Pro subscribers
  • Ranked #1 on the Vals Index and 87% on ERQA embodied reasoning (Google-reported)
  • Free 6-month top Antigravity tier offered to 10,000 scientists; Fairwind expanded to hundreds more partners

Use cases

  • Autonomous security vulnerability discovery and remediation for defenders
  • Long-horizon agentic coding and software engineering
  • Enterprise knowledge work in legal and finance analysis
  • Government and defender pre-release cybersecurity programs

Pricing

Introductory API pricing $2 per million input and $10 per million output tokens (cached input 95% cheaper) until November 30, 2026, then $4 / $20. Gemini app access requires Google AI Ultra or Pro.

Pricing reflects the source record and may change. Confirm current terms with the provider.

Dates and sources

Released
30 Sept 2026
Record updated
15 Oct 2026
Record last checked
15 Oct 2026

Explore the Feed

Other entries filed under frontier/foundation models; inclusion does not imply equivalent features.

  • Muse Wonder

    Meta's most capable Muse models, with active memory and long-goal focus for agents, 1M context and $4/$20 per million tokens.

  • Rho-1

    Reka's 19B omni-reasoning model that handles text, images, video and robot actions in one network, in research preview.

  • Amazon Nova 3

    Amazon's third-generation Nova frontier model family, unveiled at an AWS keynote with a 1M-token context window and a dedicated agentic tool-use variant.

Browse the Feed