OpenAI bets big on ‘Spud’ as next-gen AI push sharpens AGI race

OpenAI is doubling down on its next-generation artificial intelligence model, codenamed “Spud,” with President Greg Brockman describing it as the culmination of nearly two years of research and a decisive step toward artificial general intelligence (AGI).

Speaking on the Big Technology podcast, Brockman suggested the upcoming model—widely speculated to be GPT-5.5—could set a new industry benchmark, marking a significant leap over existing systems. The push comes as the OpenAI intensifies focus on its core AI offerings amid growing competition from rivals like Anthropic.

‘Big model smell’ signals smarter AI leap

Brockman hinted that Spud will introduce what he described as a “big model smell”—a term he coined to reflect a noticeable jump in intelligence, akin to the distinct feel of a brand-new product.

He emphasised that the model will significantly improve contextual understanding, reducing the need for users to repeatedly clarify prompts. The goal, he said, is to create systems that can intuitively grasp intent and deliver accurate responses with minimal input.

“When AI doesn’t quite get a question, it’s frustrating,” Brockman noted, adding that Spud aims to eliminate that gap by enabling more natural, effortless interactions. Users, he suggested, will be able to rely on the system for a wide range of tasks “without thinking very much.”

Currently in its pre-training phase, Spud is expected to roll out later this year and is seen as central to OpenAI’s strategy to streamline operations and strengthen its product ecosystem.

Strategic reset: Sora shelved as AGI takes priority

As part of this sharper focus, OpenAI has paused several non-core initiatives, including plans for an “erotic mode” for ChatGPT and the shutdown of its video-generation platform Sora.

The decision to wind down Sora—despite a reported billion-dollar partnership with Disney—underscores a broader strategic shift. According to Brockman, reallocating computational resources toward core AI development is critical to accelerating progress toward AGI.

He argued that advances in text-based AI have already demonstrated a clear trajectory toward human-like intelligence. “We have line of sight to AGI,” he said, suggesting that the long-debated limits of text models are now being surpassed.

AGI—the ability of machines to think and reason at human levels—has been a founding ambition of OpenAI since its inception. Brockman indicated that recent breakthroughs are bringing that vision closer to reality, with Spud positioned as a key milestone in that journey.

As competition intensifies and stakes rise, OpenAI’s renewed focus signals a high-risk, high-reward bet: sacrificing short-term ventures to accelerate the race toward the ultimate goal of artificial intelligence.