OpenAI Declares 'Code Red' for ChatGPT

04 Dec 2025

OpenAI has declared an internal "code red" to accelerate ChatGPT development after competitors — notably Google’s Gemini 3 and the Nano Banana Pro offering — made significant advances. According to an internal memo from CEO Sam Altman cited by The Rundown, the company is shifting priorities toward the consumer ChatGPT experience and fast‑tracking new reasoning capabilities to remain competitive.

The push includes a near‑term rollout of a reasoning model (referred to in reporting as “Shallotpeat”) that reportedly outperforms Gemini 3 on benchmarks, plus a larger model effort codenamed “Garlic” aimed at a 2026 release (potentially a GPT‑5.2/5.5‑class upgrade) to address past pre‑training shortcomings. As part of the sprint, OpenAI is said to pause or delay advertising and some AI agent initiatives to concentrate engineering resources on personalization, image generation, and core ChatGPT improvements.

The move flips a 2022 dynamic — when Google issued its own "code red" after ChatGPT’s rise — and highlights intensifying competition from Google, Anthropic, and open‑source Chinese efforts. The memo signals OpenAI’s urgency to protect its market lead by prioritizing product quality and reasoning performance over broader expansion efforts for now.

Source