Google’s new Gemini multimodal AI model looks great. But there are questions about its technical potential, given the achievements of OpenAI, reports Bloomberg.
It’s all about the comparison between Gemini Ultra and GPT-4 that Google conducted. The company compared the capabilities of the models and presented the results in a table.
According to it, Gemini Ultra is indeed superior to GPT-4 in most indicators. However, this difference is mostly minimal, even though OpenAI’s AI model was not released yesterday.
With this in mind, the publication suggests that Google’s AI model will not remain the best for long. The company has only now managed to catch up with OpenAI, which had enough time to work on GPT-5.
This week, Google unveiled Gemini, its “most advanced and versatile AI model to date,” which could be an important milestone in the development of AI. To demonstrate the model’s capabilities, Google DeepMind created a short video showing Gemini’s performance.