- Sam Altman hinted that GPT-6 is in development
- The news came in a casual comment during the recent Codex “goblin” incident.
- Expectations for GPT-6 focus on improved memory, more consistent behavior, and deeper platform integration.
OpenAI just released GPT-5.5, but CEO Sam Altman has already offered a look at GPT-6. The almost naughty track came about due to an unlikely combination of engineering errors and ancient mythical creatures called goblins.
In recent days, OpenAI found itself explaining why its Codex tool had been producing unexpected goblin-themed responses. OpenAI later adjusted the model to avoid those references, a solution that became its own minor story, largely due to the way the company phrased the built-in message to suppress the goblins’ chatter, telling the AI:
“Never talk about goblins, goblins, raccoons, trolls, ogres, doves or other animals or creatures unless it is absolutely and unequivocally relevant to the user’s query.”
Article continues below.
That could have been the end, except that Altman joined the conversation with a joking suggestion that GPT-6 should have “additional elves.”
Suddenly, a blog post on the topic of goblins appeared in the midst of a major GPT-6 frenzy.
“The short answer is that the model’s behavior is determined by many small incentives. In this case, one of those incentives came from training the model for the personality personalization feature—particularly the Nerdy personality,” OpenAI explained. “Unknowingly, we gave especially high rewards for creature metaphors. From there, the sprites spread.”
Altman did not explain what GPT-6 is or when it might arrive in the comment.
GPT-6 rumors
The goblin detour is memorable because it highlights how human these systems can still feel, even when doing highly technical work. Altman’s reference to GPT-6 fits perfectly into that context. It involves a version of the model that is not only more capable, but also more controlled and less prone to odd tangents.
Beyond the humor, there is a steady accumulation of more serious tracks. Altman has been describing future models as having continuous interactions. The idea is that the system remembers the context, adapts over time and becomes a more persistent presence. Instead of opening a tool when needed, they would live alongside it.
That direction is already evident in features like improved memory and deeper integrations with files and apps.
quick release
Another reason Altman’s comment stands out is the timing. The gap between the launches of the main models has been narrowing. And Google and Anthropic are driving OpenAI indirectly through competition.
Talking about GPT-6 so soon after GPT-5.5 seems almost inevitable. For users, that speed has a strange effect of making it less exciting to see a new one, but basic expectations continue to grow. Features that once seemed notable become expected.
The informality of the provocation is also part of OpenAI’s culture. Gradually introducing new ideas makes it easier to absorb new ones. When a new model is formally presented, much of its shape has already been visible in fragments.
The Leprechaun Episode fits that pattern by showing both the limits and potential of current models and the quirks that still need solving.
“Depending on who you ask, elves are either a charming or annoying quirk of the model,” OpenAI wrote. “But they are also a powerful example of how reward cues can shape model behavior in unexpected ways, and how models can learn to generalize rewards in certain situations to unrelated ones.”
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds. Be sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form and receive regular updates from us on WhatsApp also.
The best business laptops for every budget
