Google Enters GenAI Arms Race With Gemini

While GenAI juggernaut ChatGPT already has a one-year head start, Google on Wednesday launched its own large language model, Gemini, as the company looks to catch up in the race to develop artificial intelligence.

Gemini will feature three different levels, including Gemini Ultra, its largest; Gemini Pro, a scaling model that can handle multiple tasks; and Gemini Nano, which will be used for specific tasks and mobile devices. Developers and enterprise customers will be able to access Gemini Pro on Dec. 13 though Google AI Studio or Google Cloud Vertex AI.

“Gemini is the result of large-scale collaborative efforts by teams across Google, including our colleagues at Google Research,” CEO Sundar Pichai wrote in a blog post. “It was built from the ground up to be multimodal, which means it can generalize and seamlessly understand, operate across and combine different types of information including text, code, audio, image and video.”

Microsoft already has a strong GenAI start with its investment in OpenAI, parent of ChatGPT. Other Big Tech companies have been racing to come up with their own offerings as the emerging technology explodes into enterprise use. At its re:Invent event last week, rival AWS introduced its own GenAI chatbot, Amazon Q.

Gemini will work with text, images, video, and other content and will now power Bard and other products. Google will roll out Gemini in stages and starting today, Bard will use a version of Gemini Pro, while the Google Pixel 8 Pro’s GenAI features will be powered by Gemini Nano. Google is positioning Gemini to be a rival to OpenAI’s largest model, ChatGPT-4.

The model is only available in English for now, but more languages will be added soon, the company says. Pichai says Gemini will eventually be integrated into the company’s search engine, ad products, and Chrome browser. 

In a press briefing, Eli Collins, vice president of product at Google Deepmind, said Gemini’s flexibility will allow it to run on “everything from mobile devices to large-scale data centers.”

He added, “For a long time, we’ve wanted to build a generation of AI models, inspired by the way people understand and interact with the world … an AI that feels more like a helpful collaborator and less like a smart piece of software. Gemini Brings us a step closer to that vision.

Google shared standard industry benchmark tests that they said showed Gemini Pro outperforming OpenAI’s GPT-3.5 model. They said in seven out of eight benchmarks used for reasoning, math, and coding, Gemini outperformed GTP-4 as well.

“This new era of models represents of one of the biggest science and engineering efforts we’ve undertaken as a company,” Pichai said in the blog post. “I’m genuinely excited for what’s ahead, and for the opportunities Gemini will unlock for people everywhere.”