Altman memo: new OpenAI model coming next week, outperforming Gemini 3

OpenAI is preparing to launch a new reasoning model that, according to internal tests, outperforms Google's upcoming Gemini 3. The release is part of a broader strategic shift that will push other projects to the side.

According to an internal memo from CEO Sam Altman, reported by The Information, OpenAI plans to ship the new reasoning model as early as next week. Altman says internal evaluations place it ahead of Google's Gemini 3. It is still unclear whether this refers to the Shallotpeat model.

The accelerated launch is part of a company-wide effort internally labeled "Code Red." OpenAI is responding to increasing competitive pressure from Google, whose chatbot Gemini grew from 450 million monthly active users in July to 650 million in October. Google has also strengthened its position as Gemini 3 now leads in numerous benchmarks. Altman warned staff about "temporary economic headwinds" caused by Google's renewed surge.

Advertising and agent projects get pushed back

To focus fully on ChatGPT improvements and the upcoming model, OpenAI is delaying other commercial projects, The Information reports. Plans to introduce advertising and develop autonomous AI agents for shopping and health tasks are on hold. Development of Pulse, a service for personalized briefings, no longer has any priority.

Resources will instead go toward improving model behavior in ChatGPT and, more urgently, toward image generation. OpenAI's own image tools are under pressure after Google released its powerful new model Nano Banana Pro alongside Gemini 3.

Three years ago, the situation was reversed. Google declared its own "Code Red" to counter the sudden threat ChatGPT posed. That internal shake-up led to significant restructuring and eventually to the launch of the Gemini family.