Google hopes it has seized the AI initiative with Gemini

For most of this year the clear leader in the AI arms race was OpenAI but its recent leadership crisis potentially left the door ajar for its rivals to steal a march. The launch of the Gemini model seems to be Google’s most significant AI development for some time, maybe even ever. If its own claims are to be believed then Gemini may well have put Google back in the lead of the AI race.

Now, we’re taking the next step on our journey with Gemini, our most capable and general model yet, with state-of-the-art performance across many leading benchmarks,” said Google CEO Sundar Pichai. “This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company. I’m genuinely excited for what’s ahead, and for the opportunities Gemini will unlock for people everywhere.”

“Gemini is the result of large-scale collaborative efforts by teams across Google, including our colleagues at Google Research,” said Demis Hassabis, CEO of Google DeepMind – its AI arm. “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.

“Gemini is also our most flexible model yet — able to efficiently run on everything from data centers to mobile devices. Its state-of-the-art capabilities will significantly enhance the way developers and enterprise customers build and scale with AI.”

It’s that flexibility and scalability, rather than raw processing power, that may be the biggest differentiator for Gemini. It has been launched in three varieties: Ultra is the full fat one, while Pro is a stripped down version designed for relatively constrained computing environments, while Nano seems to be designed to run on mobile devices. Here’s how Google reckons Gemini Ultra compares to OpenAI’s best model – GPT4 – against a bunch of benchmarks.

The initial response seems to be fairly positive, with MIT Technology Review saying it “looks amazing”, while warning against excessive AI hype. TechCrunch, meanwhile, seems to think it falls a bit short of the expectations Google set in the build up to the launch. Gemini Pro has been plugged into Bard, so you can make your own mind up by giving it a whirl. Here’s a question we asked with the first bit of the response.

The popularity of OpenAI’s ChatGPT was arguably an existential threat to Google as it provided a superior search alternative. Now that OpenAI is apparently more directly controlled by Microsoft that threat is bound to grow, but its recent dramas must surely have slowed its progress so this is the ideal time for Google to pounce. Meanwhile the other main hyperscaler, AWS, is still struggling to catch up if reports from its recent re:Invent event are anything to go by.

Here's Google's video announcing the launch of Gemini.