OpenAI CEO Sam Altman’s Bold Statement on GPT-5
In a recent panel discussion at TU Berlin, OpenAI CEO Sam Altman made a striking comment about the future of artificial intelligence. He stated, “I don’t think I’m going to be smarter than GPT-5.” His remark highlights the rapid advancements in AI technology and the growing capabilities of OpenAI’s next-generation model.
Challenging the Audience
During the discussion on “The Age of AI,” Altman posed a thought-provoking question to the audience:
“How many of you think you’re still going to be smarter than GPT-5?”
This question underscores the accelerating progress in AI, hinting that GPT-5 might surpass human intelligence in specific domains.
Altman went on to explain that he doesn’t view this development as a negative. Instead, he emphasized the potential for AI to drive scientific breakthroughs and enable researchers to achieve previously unimaginable feats. “We want more science to get done. We want to enable researchers to do things they couldn’t do before,” he said.
Is GPT-5 Almost Ready?
Altman’s comments suggest that OpenAI is nearing the completion of training for GPT-5. While the company has not officially announced a release date, there are signs that post-training evaluations indicate significant advancements in AI capabilities.
Despite this, Altman mentioned in a recent Reddit AMA that there is still no official timeline for GPT-5’s release. However, he has been speaking more openly about the model in recent interviews, fueling speculation about its imminent arrival.
Breakthroughs in the O-Series Models
While the AI community eagerly anticipates GPT-5, OpenAI has been focusing on its o-series reasoning models. The o3 model, in particular, has made headlines for passing the ARC-AGI benchmark—a critical test for artificial general intelligence (AGI). However, this achievement came at a significantly higher computational cost.
One of the most impressive feats of the o3 model is its performance in Deep Research agent tasks. With access to web search and Python tools, it scored 26.6% on Humanity’s Last Exam, a benchmark for AI problem-solving. In contrast, competing models have oI is nearing the completion of training for GPT-5. While the company has not officially announced a release date, there are signs that post-training evaluations indicate significant advancements in AI capabilities.
Despite this, Altman mentioned in a recent Reddit AMA that there is still no official timeline for GPT-5’s release. However, he has been speaking more openly about the model in recent interviews, fueling speculation about its imminent arrival. interviews, fueling speculation about its imminent arrival.nly reached around 9.4% without web search capabilities
The Future of AI: Merging GPT-Series with O-Series
During his visit to Japan, Altman revealed that OpenAI aims to integrate the GPT-series and o-series models into a single, highly capable AGI system. “At some point, the GPT series will converge with the o-series to create an inference-scaled model that is far more capable and intelligent,” he explained.
This ambitious goal could mark a turning point in AI development, bringing humanity closer to artificial general intelligence—machines that can perform any intellectual task a human can.
Final Thoughts
Altman’s statements at TU Berlin reinforce the idea that AI is evolving at an unprecedented pace. Whether GPT-5 will truly surpass human intelligence remains to be seen, but one thing is clear—OpenAI is pushing the boundaries of what AI can achieve. With plans to integrate GPT and o-series models, the future of AI looks more transformative than ever.
Stay tuned for more updates on the latest advancements in AI and technology at Malwerm Tech!
