For Immediate IT Support 613-288-5805 or email us

Google Announces Gemini – A Powerful GenAI Model for All Devices

Google catapulted us towards a new world of artificial intelligence by launching Gemini, an immensely powerful generative AI (genAI) model. Ranging in three different sizes, this flexible AI model can be utilized in various spaces—from data centers to mobile devices[1%5E].

A Leap Forward in AI Capabilities

Over the past eight months, Google has been meticulously developing the Gemini large language model (LLM), recently making it accessible for a select group of companies for initial use[1%5E].

This conversational genAI software presents itself as Google’s most powerful tool so far and has immense potential to compete with other advanced LLMs such as Meta’s Llama 2 and OpenAI’s GPT-4[1%5E].

Unveiling this AI model is a significant milestone for Google. “This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company,” wrote Google CEO Sundar Pichai[1%5E].

Delving Deeper into Gemini’s Power

Gemini is a multimodal model that can handle multiple methods of input, such as photos, audio, and video. Unlike other multimodal models which require different components for different modalities, Gemini is designed to be natively multimodal, being pre-trained on different modalities and further refined with additional multimodal data[1%5E].

The three versions of Gemini 1.0 cater to different potential applications[1%5E]:

  • Gemini Ultra, the largest model, is designed for handling highly complex tasks.
  • Gemini Pro is best suited for day-to-day tasks across a broad range.
  • Gemini Nano is created specifically for on-device tasks.

Additionally, Google introduced its powerful ASIC chip — the Cloud TPU v5p — tailored for the massive processing demands of AI[1%5E].

Gemini Across Google’s Core Products

Google’s new Gemini LLM is already making its presence felt in some of Google’s core products, such as the Bard chatbot using a version of Gemini Pro for advanced reasoning and understanding. Google’s Pixel 8 Pro is engineered for Gemini Nano, incorporating it in features like Summarize in Recorder and Smart Reply in Gboard[1%5E].

In the coming months, Gemini powered features are set to appear in more of Google’s products like Ads, Chrome, and Duet AI[1%5E].

Developers keen to build Gemini-powered apps for mobile devices can sign up for an early preview of Gemini Nano via Android AICore[1%5E]. On another note, from Dec. 13, developers and enterprise customers can access Gemini Pro via the Gemini API in Vertex AI or Google AI Studio[1%5E].

Stepping Up Processing Capacity

In conjunction with Gemini, Google has ramped up its processing capacity with its AI Cloud Hypercomputer — an integrated system of performance-optimized hardware, open software, machine-learning frameworks, and flexible consumption models, offering improved efficiency and productivity across AI training, tuning, and serving[1%5E].

The launch of Gemini embodies the exciting new era of AI models and the constant pursuit of improvement and innovation.



  1. ComputerWorld: Google launches Gemini, a genAI model for all devices