Google DeepMind has officially introduced its highly anticipated family of AI chatbots named Gemini, poised to compete with OpenAI’s renowned GPT series.
Among the key highlights is Gemini Ultra, the largest and most advanced model, which Google claims surpasses OpenAI’s GPT-4 in various domains, including text-based, image-based, coding, and reasoning tasks.
The announcement also shed light on the meticulous development process, emphasizing that Gemini Ultra is undergoing rigorous “trust and safety checks, including red-teaming by trusted external parties.”
This stringent evaluation process aligns with Google’s commitment to ensuring the reliability and security of its AI technologies.
Accompanying Gemini Ultra are two additional models, Gemini Pro and Gemini Nano. Gemini Pro is now accessible to the public through Google’s Bard chat interface, while the smaller Gemini Nano is designed to run on Google’s Pixel 8 Pro smartphone.
All three models exhibit the capability to process text, images, audio, and video, providing comprehensive outputs in both text and image formats.
Google envisions the integration of Gemini models into various products and services, including internet search and advertisements.
Developers will gain access to Gemini Pro through an API starting December 13, with Android developers empowered to build with Gemini Nano.
The Gemini suite is set to face competition from industry rivals, including OpenAI, Anthropic, Inflection, Meta, and Elon Musk’s xAI.
Google DeepMind’s ambitious move reflects its dedication to advancing AI capabilities and establishing a strong presence in the burgeoning field of AI chatbots.
This unveiling marks a significant milestone for Google DeepMind, a company born out of the fusion of DeepMind and Google Brain in April 2023.
The incorporation of Gemini into Google’s AI portfolio signifies the tech giant’s determination to close the gap with competitors and assert itself as a leader in AI innovation.