Particle.news

Google Releases Gemini 4 Argon for Long Technical Workflows

Global developers and enterprises can access the model via AI Studio, the Gemini API or select Google AI subscriptions under a limited introductory pricing plan.

Overview

  • Gemini 4 Argon is now publicly available to developers and enterprises through AI Studio, the Gemini API and designated Google AI subscription tiers.
  • Google is offering time‑limited introductory per‑million‑token prices for input, output and cached input that rise after the promotion ends on November 30, 2026.
  • The model supports a shared context and decoding capacity of 1,048,576 tokens so input, internal reasoning and output compete within the same one‑million‑token window.
  • Google says Argon leads the Vals Index on complex tasks across code, finance, legal, security and science and cites internal case studies such as a Chrome codebase refactor and open‑source performance gains.
  • Google reports multi‑layer safety controls, an expanded Fairwind security program and a six‑month Antigravity Ultra trial for 10,000 scientists, while noting independent verification and long‑term operational experience remain to be seen.