2 min read

A tiny web campaign wants GPT-5.5 kept alive

A personal site urges OpenAI to preserve GPT-5.5, arguing the model and its chat harness are worth keeping available long-term.

Image: Hacker News

A small website titled “Save GPT-5.5” is making a straightforward plea: don’t retire GPT-5.5.

The post, shared on Hacker News and hosted at save-gpt-5-5.fyi, is less a formal campaign than a deeply personal endorsement of OpenAI’s GPT-5.5, specifically “chatgpt-5.5-thinking” on chatgpt.com and the Codex desktop app. The author says the model delivers a “mostly satisfying experience” and argues that this particular combination of model and product interface is worth preserving.

Rather than framing the issue as a step toward artificial general intelligence, the author describes it as a case for archiving a useful tool while public demand still exists. They suggest that keeping an older model online could make economic sense because older systems are typically cheaper to run, and because strong user attachment has public-relations value for OpenAI.

The post also doubles as a call for internet-style grassroots pressure. Readers are asked to spread the message, self-host copies, and build support for keeping GPT-5.5 available “forever,” or at least well beyond a normal retirement cycle.

The writing is intentionally messy, comic, and self-aware. The author insists no writing assistance was used for the plea itself, though they say they asked GPT-5.5 to help create the site’s simple static HTML. They also note that “5.6 already out”, giving the appeal a note of urgency.

Recommended reading

Some AI problems stay unsolvable, even with infinite data

There is no indication in the post that OpenAI plans any response. For now, the campaign is simply a user’s public attempt to turn affection for a specific model version into pressure to keep it around.

Ava Chen

AI Editor

Ava covers the rapidly evolving world of artificial intelligence, from foundational models and research labs to the real-world economics of intelligence. With a background in computational linguistics, she cuts through the hype to find out what actually works. She firmly believes that benchmarks are just marketing until reproduced in the wild.

via Hacker News

// Keep reading