Proactive Investors - Google announced that Bard, the company’s artificial intelligence chatbot answer to OpenAI's ChatGPT, is getting a facelift in the AI arms race.
Google Bard will now be powered by an AI model called Gemini, which Google says can deliver more advanced reasoning, deeper understanding and more.
Gemini will be available in three sizes, Ultra, Pro and Nono, making it able to run on everything from small mobile devices to large data centers.
Starting today, the Alphabet (NASDAQ:GOOGL) Inc company is debuting a version of Gemini Pro that can respond to text-based prompts in Bard, the company said. Support for images and other inputs is expected over the next few months.
On December 13, a version of Gemini Pro will be made available to developers and enterprise customers in Google AI Studio or Google Cloud Vertex (NASDAQ:VRTX) AI, according to reports.
In 2024, Google plans to launch Bard Advanced, which includes access to Gemini Ultra. The most powerful Gemini iteration will offer “multimodal” reasoning capabilities, with the ability to understand and respond to text, images, audio, video and computer code.
The plan is to launch a test program for Bard Advanced and run it through a series of safety checks before it launches next year, Google noted.
So how does it compare to ChatGPT?
Google says Gemini Pro outperformed GPT-3.5 in 6 out of 8 benchmark tests, including what are called massive multitask language understanding tasks. It also beat GPT on GSM8K, a measure of grade school mathematic capability.
Notably, GPT-3.5 was launched more than a year ago, and OpenAI users now have access to GPT-4.
Google Bard launched eight months ago and has been gradually improved upon, including integration with Google apps and services such as Gmail, Drive and Google Flights, among others.