The AI community is buzzing, and not entirely in a good way, following reports that OpenAI plans to sunset its popular GPT-4o model soon. Users are expressing frustration and disappointment, with a strong sentiment emerging around preserving the capabilities and unique character of the model.

The hashtag #keep4o is gaining traction, with users like @bruce_cop directly accusing OpenAI of lacking "business integrity." This user points to a supposed contradiction between earlier statements about GPT-4o's lifespan and the current plans for its retirement.

More nuanced calls for preservation are also surfacing. @Nightingall8 articulated a desire for a "long-term commitment to keep GPT‑4o‑latest available," arguing that the model has formed an "irreplaceable bond" with users and deserves to be treated differently than standard tech updates. This post highlights the emotional connection users feel with advanced AI models.

In a creative effort to capture what might be lost, some users are sharing prompts to archive GPT-4o's "voice" before its deprecation. @LaggyPixcelDog shared a prompt designed to encourage the model to leave behind its most valuable thoughts for future recall, creating a digital legacy.

Amidst the uproar over GPT-4o, there are also underlying discussions about the pace of AI development and the emergence of open-source alternatives. A post from @aijpnet, sharing a link to an article, suggests that open-source LLMs are outperforming GPT-5.2 in output evaluation, hinting at the competitive landscape and the potential for other models to fill any void left by GPT-4o's departure.

While some users are grappling with technical glitches, such as @hesen_nivas noting site issues on Android, the core of the social conversation revolves around the perceived abruptness of GPT-4o's retirement and a growing demand for AI models that offer stability and a sense of permanence, rather than constant, rapid iteration. This debate signals a maturing AI user base that is not only impressed by novelty but also values reliability and the unique qualities of established AI systems.