Is GPT-5.5 “Spud” real? Official sources still don’t confirm a release

That does not prove Spud does not exist internally. It simply means the public evidence does not support writing that OpenAI has formally launched “GPT-5.5 Spud.”

What OpenAI’s official pages currently show

The verifiable OpenAI material points to GPT-5 and GPT-5.4, not GPT-5.5:

• GPT-5 has official OpenAI pages. OpenAI has a GPT-5 introduction page and a GPT-5 product page with links for ChatGPT users and developers.
• The API documentation shows GPT-5 and a GPT-5.4 signal. OpenAI’s API docs include a GPT-5 model page, and the documentation navigation labels the latest model guide as Latest: GPT-5.4.
• Formal API models usually have their own documentation. In the provided sources, GPT-5 nano and GPT-5-Codex each have API model pages. No equivalent GPT-5.5 or Spud model page appears in the supplied official documentation.
• GPT-5.4 has a clearer public rollout trail. An OpenAI Developer Community post in the Announcements category says GPT-5.4 Pro and Thinking are here and that GPT-5.4 is gradually rolling out to ChatGPT and Codex.

Those sources make GPT-5 and GPT-5.4 verifiable. They do not confirm that Spud is GPT-5.5, nor do they provide a GPT-5.5 model ID.

Where the Spud rumor comes from

The public material connecting “Spud” with GPT-5.5 is largely framed as a leak, rumor or speculation. Geeky Gadgets uses the phrase “OpenAI Spud Leak” in its headline, while India Today describes Spud as “rumoured to be GPT 5.5.”

TokenMix’s GPT-5.5/Spud article is also cautious in an important way: its snippet says there is no official GPT-5.5 release date, no model card and no API pricing, and it labels some of its projections as speculation.

Those articles can show that the AI community is discussing Spud. They do not replace an OpenAI product announcement, API documentation or model card.

OpenAI community posts are not the same as launch notices

There are SPUD-related posts on the OpenAI Developer Community, but the examples in the provided sources appear under feature-request or ChatGPT discussion contexts. They ask for an optional expression mode or refer to a hoped-for “SPUD release.”

That is different from an official Announcements post such as the GPT-5.4 Pro and Thinking rollout. A forum post appearing on OpenAI’s community site does not, by itself, make a term an official ChatGPT or API model name.

How to tell if an OpenAI model has really launched

For future model rumors, the strongest signals are usually:

1. An OpenAI News or product-release page. OpenAI’s News page provides official categories for product releases, research, safety, engineering and company updates.
2. API model documentation. OpenAI’s developer docs list model pages and latest-model guidance, such as the GPT-5 model page and the Latest: GPT-5.4 guide signal.
3. A system card or model card. OpenAI has published GPT-5 system-card updates, including for GPT-5.2; a major GPT-5.x release would be more credible if accompanied by comparable official safety or model documentation.
4. A model ID, pricing and availability details. Official model information commonly includes developer entry points, model pages, usage details or pricing context. The provided sources do not include those items for GPT-5.5 or Spud.

The safest wording for now

A careful description would be: “Spud may be an external rumor or codename, but as of early May 2026 there is not enough public evidence to show that OpenAI has officially released GPT-5.5 ‘Spud.’”

Avoid saying: “OpenAI has launched GPT-5.5 Spud.” That claim currently lacks the expected support from an OpenAI announcement, API model page, model card, pricing page or availability documentation. The available evidence supports a more limited conclusion: GPT-5 and GPT-5.4 are publicly verifiable; GPT-5.5 “Spud” is not confirmed as an official release in the provided sources.

In short, Spud may still be a real codename, a future model or simply a rumor. But until OpenAI publishes official documentation, developers, journalists and content creators should label it as unconfirmed rather than as a released model.