Google releases Gemini 4 Argon globally with agentic coding claims and cost schedule

Google released Gemini 4 Argon globally, positioning it for coding, knowledge work, enterprise workflows, and cyber defense. Access spans the paid Gemini API and Gemini platform; Google AI Pro and Ultra in the Gemini app and AI Mode in Search; and Ultra in Antigravity. The rollout supersedes an earlier claim that access was restricted to cybersecurity defenders.

Introductory API pricing is set at $2 per million input tokens and $10 per million output tokens, with cached input discounted 95%. Artificial Analysis reports that rates double to $4 per million input tokens and $20 per million output tokens after November 30, a qualification relevant for cost comparisons.

Google describes internal deployment outcomes, reporting that Argon helped separate core webpage handling from a six-million-line legacy Chromium folder, completing months of work in weeks and eliminating a category of browser crashes. It also reports a 35% increase in agentic code submissions and a nearly one-third reduction in rollbacks. The company notes these are internal deployment claims, not controlled productivity measurements.

Argon is available in developer tools including Cursor, Devin Desktop/CLI, and GitHub Copilot. Cursor reports Argon at 52.8% on CursorBench at default effort, on par with Opus 5.5. Cognition reports 60.5% on FrontierCode 1.1 and a 75% Devin discount through October 27.

Artificial Analysis estimates Argon’s cost per task at approximately $1.99 under introductory pricing, placing it level with GPT-6 Astra on its Intelligence Index at roughly 60% of Astra’s cost per task. It also reports substantial agentic gains and a 15% AA-Omniscience hallucination rate, the lowest among models scoring at least 45 on the index.

Physical-task demonstrations include DeepMind researchers describing Argon deriving a cornhole simulation from human video and refining throws on Apptronik’s Apollo robot, achieving seven consecutive successful throws. The described workflow included real-robot attempts and refinement, despite references to a zero-shot setup.

Argon’s agent capabilities are reflected in rankings such as Arena, where it is placed High fifth behind Opus 5.5, GPT-6.1 Sol, Fable 5.1, and GPT-6 Astra at Max effort. Its revised Bash Reliability signal isolates agent-caused failures and credits correct first-time command use.

Google is offering 10,000 scientists Antigravity’s highest usage tier free for six months through an application program, and cybersecurity defenders can apply through Fairwind. Google also highlights Antigravity’s coding interface entering the Gemini macOS app and connections to more than 40 apps, including Instacart, OpenTable, Adobe, and Shopify.