Three model names that OpenAI has not announced, aurora-red-alpha, aurora-green-alpha and aurora-blue-alpha, have been reported in the OpenAI API. The AI writer Andrew Curran posted on X at 03:13 UTC on 18 October that “Three mysterious new models just showed up in the OpenAI API”.
Crypto Briefing reported the same three identifiers a few minutes later, writing that OpenAI “has not announced them, documented them, or explained what they do.” When we checked OpenAI’s developer changelog and models page at about 11:35 UTC, neither page mentioned an aurora model.
Business Pill · AI EVALUATION
A short explainer on treating AI evaluation as an acceptance test: how to judge whether a model is good enough before relying on it. It teaches the general idea only and says nothing about OpenAI or the aurora models.
The key insight: As we read it, the three names are a signal without a spec: Curran’s post and Crypto Briefing’s report give the IDs, and Crypto Briefing says there is “no public pricing and no stated use case.”
Curran’s post lists the three names and nothing else:
“Three mysterious new models just showed up in the OpenAI API: aurora-red-alpha aurora-green-alpha aurora-blue-alpha”
Crypto Briefing’s report says the three IDs are “now listed in OpenAI’s developer interface” and that “No changelog entry confirms their status. There is also no public pricing and no stated use case.” It traces the wider attention to a post on X by the AI researcher @scaling01.
Andrew Curran posted on X at 03:13 UTC on 18 October that three new models, aurora-red-alpha, aurora-green-alpha and aurora-blue-alpha, had shown up in the OpenAI API; Crypto Briefing reported them at 03:18 UTC with no documentation or pricing. Neither OpenAIโs developer changelog nor its models page mentioned an aurora model when we checked at about 11:35 UTC. Sources: Andrew Curran on X; Crypto Briefing; FourWeekMBA check of OpenAIโs pages.
What Was Posted on X
Three mysterious new models just showed up in the OpenAI API:
Andrew Curran’s post of 03:13 UTC on 18 October lists the three model names, aurora-red-alpha, aurora-green-alpha and aurora-blue-alpha, and says they showed up in the OpenAI API. It is the first public mention we found and the basis for this piece; it gives no pricing or capabilities.
What We Could and Could Not Check
We searched OpenAI’s developer changelog and its models page for “aurora” at about 11:35 UTC on 18 October and found no match on either. We do not hold an OpenAI API key, so we could not query the model list ourselves, and we did not contact OpenAI.
So the listing itself rests on Curran’s post and Crypto Briefing’s report. Neither gives pricing, context length or capabilities for the three models.
The Earlier Aurora Alpha
The name has appeared before. OpenRouter, which routes developer requests across models from many providers, lists a model called Aurora Alpha and says it “was released on February 9, 2026.” Its page describes it as “a cloaked model provided to the community to gather feedback. A reasoning model designed for speed.”
OpenRouter’s page does not name the company behind Aurora Alpha. Crypto Briefing writes that a link to a GPT-5.3-Codex-Spark variant has been speculated but “has not been confirmed by OpenAI.”
OpenRouter’s Aurora Alpha, per its model page. Its page does not name the company behind it, and it is not established that it is related to the three IDs reported in OpenAI’s API. Source: OpenRouter.
How Cloaked Models Have Surfaced Before
OpenRouter has run other unnamed test models that were later identified. On the same page it lists Polaris Alpha, which it now says “was an early snapshot of GPT-5.1 with reasoning effort set to minimal”, and Sherlock Think Alpha, which it says “was an early snapshot of Grok 4.1 Fast with reasoning enabled.”
Those are OpenRouter’s own labels for its cloaked models. They do not tell us what the three aurora names reported in OpenAI’s API are.
The Structural Read
As we read it, the three IDs share one name and differ only by colour, and neither source says what separates them, and OpenAI’s changelog and models page did not mention them when we checked.
As we read it, the precedent is OpenRouter’s cloaked models: Aurora Alpha was released there on February 9, 2026 without a named provider, and OpenRouter later labelled other cloaked models, such as Polaris Alpha, as early snapshots of named releases.
As we read it, the useful question for builders is what an alpha ID implies about access, since Crypto Briefing reports no pricing and no documentation for any of the three.
Andrew Curran, on X, 18 October 2026
“Three mysterious new models just showed up in the OpenAI API”
Three Implications
DEVELOPERS ON OPENAI’S API Crypto Briefing reports no documentation, pricing or stated use case for the three IDs, and OpenAI’s changelog did not list them when we checked.
TEAMS WATCHING OPENROUTER OpenRouter’s Aurora Alpha page describes a cloaked reasoning model designed for speed with a 128,000-token context window, released on February 9, 2026.
ANYONE TRACKING OPENAI RELEASES Crypto Briefing says the speculation linking aurora to Codex features or faster inference has not been confirmed by anyone outside OpenAI.
The framework opens: “In any product cycle, there are two curves to follow. The first is what your underlying technology can do. The second is what your users can actually do with it.”
As we read it, an undocumented model ID sits between those two curves: Curran and Crypto Briefing report it in the API, but nothing has been surfaced to users about what it does. The framework was published in June 2026 and does not discuss aurora.
What Is Not Established
It is not established what the three models do, whether they are related to OpenRouter’s Aurora Alpha, what they cost, or whether OpenAI plans to release them. Crypto Briefing says speculation about faster inference or new Codex features has not been confirmed by anyone outside OpenAI; we have not seen any OpenAI statement either way.
Business Engineer Framework
The Product Overhang Doctrine
A Business Engineer framework on the gap between what a model can do and what a product surfaces.
Three unannounced model IDs, aurora-red-alpha, aurora-green-alpha and aurora-blue-alpha, have been reported in OpenAI’s API by Andrew Curran and Crypto Briefing. OpenAI’s changelog and models page did not mention them when we checked on 18 October.
A note on sourcing. This piece is based on Andrew Curran’s post on X of 18 October 2026, Crypto Briefing’s report of the same day, OpenRouter’s Aurora Alpha model page, and our own check of OpenAI’s developer changelog and models page at about 11:35 UTC. We do not hold an OpenAI API key and could not confirm the listing ourselves, and we did not contact OpenAI. The references to our earlier reporting link to our own pieces.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur, HackerNoon and Search Engine People. Find him on LinkedIn and Substack. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
