The force behind it was the release of competent models from Google and Anthropic, making it look like OpenAI was losing the edge in the AI race.
It was put in place last week and signaled a shift of resources from the likes of Sora and early ads development to bolstering the teams working on ChatGPT.
Not just the one model
The WSJ writes today that we can expect two models from OpenAI in the near future as a result of this.
The first one will be GPT-5.2, a stopgap to beat Gemini 3 on business and coding use, and stop the gap.
Then there will be a second one in January — with better images to counter Nano Banana, «improved speed, and a better personality,» the WSJ writes, citing sources.
OpenAI hasn’t upgraded its imaging capabilities since March.
Read the full story at The Wall Street Journal. It’s long.