Google delays announcement of "GPT-4 competitor"

The InformationAlphabet CEO Sundar Pichai has asked to delay the launch of its major language model Gemini until next year, two people familiar with the matter said. The company had previously planned to announce the new product this week at events in California, New York and Washington.

Gemini is a multimodal AI model developed by Google DeepMind, first mentioned in May at the Google I/O 2023 event. This is considered the company's "most serious effort" in the generative AI craze. Sergey Brin, Google's co-founder, also had to return to the company to promote the development of Gemini after many years of "retirement".

Gemini is Google's ambitious project and works similarly to the existing GPT-4 model on ChatGPT. However, sources say that, except for English, Gemini still performs poorly in other languages. "Google engineers found that Gemini lags behind GPT-4 in multilingual recognition and response, so there is room for improvement," a source revealed.

Previously, Google said Gemini "has impressive multimodal information processing capabilities that are unprecedented in previous models, highly efficient in integrating into tools and application programming interfaces (APIs), opening up future innovations." Google also plans to create multiple versions of Gemini to bring to its services such as Bard, Search and Workspace as well as a mobile version called Gecko.

According toInteresting EngineeringGemini could be one of Google's most important products when it launches. The model will show that the company can compete with and surpass OpenAI, as search trends shift from traditional platforms to chatbots.

In fact, no company has yet deployed an AI model that can approach OpenAI’s GPT-4. Instead, the market is flooded with products that are equivalent to GPT-3.5. In August, Dylan Patel and Daniel Nishball of research firm SemiAnalysis argued that Google Gemini “will break GPT-4 by five times the power, using fewer GPUs.”

Last month, OpenAI released an upgrade called GPT-4 Turbo that is faster and more cost-effective. In mid-November, CEO Sam Altman also revealed that the groundbreaking GPT-5 model will be introduced early next year.