Google Gemini, the company’s upcoming AI model will have multimodal capabilities, allowing it to work with various types of data like text, images and graphs.
At the Google I/O 2023, Google announced that it was working on its next-generation AI model called ‘Gemini’. The tech giant was to unveil its GPT-4 competitor sometime next week, but a new report suggests that Google might be pushing the launch date to January 2024.
According to The Information citing sources, Gemini was struggling with non-English queries, which prompted Google CEO Sundar Pichai to push back the release date to January next year. The report goes on to say that support for various languages is really important for Google to match or even surpass OpenAI’s GPT-4 and that Google has “met that standard in some respects.”
Last month Pichai said that the company is focusing on getting out Gemini 1.0 as soon as possible but is still finalising the largest version of the next-generation AI model.
What is Gemini?
When the company announced Gemini at Google I/O in May this year, the tech giant made it clear that it would be the most powerful model it has worked on. Based on PaLM 2 and led by the Brain Team and DeepMind, Google says Gemini will be ‘multimodal’, meaning it will be able to work with different types of content such as images and text and have memory and planning capabilities.
Gemini might combine various AI models and will be available in various sizes and capabilities with Google using it on all their products internally as well as roll it out to both developers and cloud customers. Pichai also added that the next-generation AI model will be “highly efficient with tools and API integrations.”
The upcoming AI model is expected to power Google’s existing AI-powered products like Search, Google Assistant and the popular chatbot Bard.
Source:indianexpress.com