DeepMind Co-Founder Reveals OpenAI Secretly Developing GPT-5, Which Will Be 100 Times Larger Than Current Models

In a recent interview, DeepMind co-founder Mustafa Suleyman revealed a shocking message to the world:OpenAIis secretly developing a product calledGPT-5new type ofartificial intelligence (AI)Model. According to him, this new model will be 1,000 times larger than the current GPT-4, and this will be realized within the next three years.

Suleyman further explained that their company, Inflection AIhas mastered an advanced training method that allows their model to be 100 times larger than the current cutting edge model within the next 18 months. He even predicts that by the next 3 years, Inflection's model will be 1000 times larger than the current version.

However, the news has sparked some controversy. Previously.OpenAISam Altman, the founder of OpenAI, has denied that he is working on GPT-5. In response, some have speculated that OpenAI may have come up with a new name for the new model, hence their denial that they were working on GPT-5. This is reminiscent of the Code Interpreter, which many believed at the time to have surpassed the capabilities of the GPT-4 model, which would have been GPT-4.5.

In this interview, Suleyman also shares some inside information about his time at DeepMind and Inflection AI, including the process of Google's acquisition of DeepMind and the issues that followed. This information goes some way to explaining why DeepMind was "up early and late" compared to OpenAI.

Additionally, Suleyman expresses concern about the problems that open source models may bring. He believes that open source models may increase the instability and harm that AI brings to humans. And he believes that the biggest threat to AI security does not come from large language models, but from autonomous intelligences that may emerge in the future.