Google has unveiled its most powerful AI model yet, called Gemini, to take on chatbot sensation ChatGPT and other rivals in the explosively growing generative AI space.
Unveiled at Google’s annual I/O conference in May, Gemini comes in three sizes – Ultra, Pro, and Nano – each suited for specific use cases in Google’s products and cloud services.
Built from scratch to understand and process text, images, video, audio, and code seamlessly, Gemini aims to power futuristic Google Search, ads, Pixel phones, and more with its unprecedented multimodal intelligence.
“Gemini represents the best of Google’s AI research to create the most capable, general, and useful model we have ever made,” said DeepMind CEO Demis Hassabis in the announcement.
While still prone to potential biases and hallucinations like other models, Hassabis notes Gemini’s comprehension and reasoning improve as its real-world knowledge expands over time.
From December 13, developers can access Gemini Pro’s text capabilities via Google’s Generative AI tools and Cloud platform to build their prototypes.
Currently English-only, multi-lingual support is incoming. Wider public testing of Gemini Ultra for feedback starts next year before full-scale release.
