A review of the company’s latest public statements, research material and remarks by senior executives shows no declaration setting December 2026 as a deadline for reaching AGI. OpenAI continues to describe the timing of artificial general intelligence as uncertain while pursuing increasingly capable systems that can perform complex professional, scientific and technical work. Its charter explicitly says the timeline to AGI remains uncertain.
The distinction is significant because OpenAI has outlined ambitious milestones for AI development without equating those targets with full AGI. The company defines AGI broadly as highly autonomous systems capable of outperforming humans at most economically valuable work, while its research material uses the concept to describe systems able to solve human-level problems across a broad range of domains.
One of OpenAI’s clearest published timelines concerns automated scientific and AI research rather than AGI itself. Chief executive Sam Altman and chief scientist Jakub Pachocki said in June that the company believes a significant fraction of its own research could be conducted by AI systems working alongside human researchers by March 2028. That programme is designed to create an automated AI researcher capable of accelerating and progressively automating parts of the research process.
OpenAI has separately said it expects AI systems during 2026 to begin making small scientific discoveries. More substantial autonomous discoveries are anticipated from 2028 onwards, although the company has acknowledged that these forecasts carry considerable uncertainty. Those targets indicate rapid advances in machine reasoning and research assistance, but they are not equivalent to an announcement that AGI will arrive this year.
The confusion reflects a broader problem surrounding the term AGI. Technology companies, researchers and investors often use different benchmarks for deciding when general intelligence has been reached. Some definitions emphasise performance across intellectual tasks, while others focus on autonomy, economic productivity, scientific discovery or an AI system’s ability to improve subsequent generations of artificial intelligence.
OpenAI’s commercial arrangements with Microsoft provide an additional formal process. Their partnership specifies that any declaration by OpenAI that AGI has been achieved must be verified by an independent expert panel. The two companies reaffirmed this year that the contractual AGI definition and the process for determining whether the threshold has been crossed remain unchanged.
That mechanism makes a casual declaration of AGI materially different from predictions made by executives, researchers or prominent technology entrepreneurs. Industry forecasts remain exceptionally varied. Some technology leaders have suggested that machines exceeding individual human intelligence could appear within a short period, while other researchers argue that existing large-language-model architectures still face obstacles in reliability, continual learning, long-horizon planning and genuinely autonomous reasoning.
OpenAI nevertheless continues to push capability boundaries rapidly. Its GPT-5.6 family, released in July, was positioned as a major advance in coding, knowledge work, cybersecurity and science. The company has also expanded access to frontier models for academic researchers and said AI-assisted research could become an increasingly important driver of scientific progress. None of those announcements characterised GPT-5.6 or another publicly available model as AGI.
Safety considerations are becoming more prominent as capabilities increase. OpenAI said on August 18 that preliminary testing of an upcoming model, Astra, suggested it could approach a critical cybersecurity capability threshold under the company’s Preparedness Framework. The finding prompted additional work on monitoring, alignment and containment, illustrating how frontier models can acquire powerful specialised abilities without necessarily satisfying a broader definition of general intelligence.
Altman has maintained an aggressive long-term outlook. Writing about OpenAI’s first decade, he said he believed the company was almost certain to build superintelligence within another 10 years. That statement points towards confidence in continued rapid progress but stops well short of promising AGI during 2026.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
