AI Roundup: Gemini 4 Argon, OpenAI Security Arrest, and Coding Agent Data

The October 14 AI roundup leads with Google DeepMind's Gemini 4 Argon, a frontier model aimed at software engineering, enterprise knowledge work, and cybersecurity defense. It also covers the arrest of an OpenAI engineer accused of hiding a recording device in an executive conference room, the collapse of an AI entertainment startup, and new findings on AI coding agents' productivity. Gemini 4 Argon is described as having a 1 million output token limit, top benchmark results, and an initial pricing plan before wider release.

Google's Argon lifts the output ceiling from 64,000 to one million tokens and reports leading scores on DeepSWE v1.1, the Vals Index, and Zapier's AutomationBench. Google says internal agents converted C/C++ code to Rust and recovered over 300 TiB of memory. Early access goes to cyber defenders via Fairwind and Wiz, with promotional token prices before standard rates.

OpenAI's Abdulrahman Alfozan was arrested after an encrypted recorder was found September 30 behind a conference-room TV, wrapped in toilet paper and connected by USB. He was fired and faces up to five years and a $250,000 fine. Character.AI entered Chapter 11 with about 12 million users, never profitable and spending $8–9 million monthly; Disney is bidding for its IP.