Gemini Pro will support text-based prompts in Bard at present with multimodal support in the coming months
Google's artificial intelligence (AI) chatbot Bard is getting its biggest upgrade yet with the launch of Gemini, the company's newest and the most advanced AI model.
Gemini will be available in three sizes: Ultra (for highly complex tasks), Pro (for scaling across a broad range of tasks) and Nano (on-device tasks).
Sissie Hsiao, vice-president of Google Assistant and Bard, said that Gemini is being rolled out to Bard in two phases: Bard will be powered by a specifically tuned version of Gemini Pro starting December 6, making the chatbot more capable of understanding and summarising, reasoning, coding, and planning.
Gemini Pro will support text-based prompts in Bard at present with multimodal support in the coming months. It will initially be available in English across more than 170 countries and territories with plans to add support for more languages and geographies such as Europe in the near future.
Early next year, Google will introduce Bard Advanced, which will give users access to the tech giant's most advanced models and capabilities starting with Gemini Ultra.
Read: Google unveils Gemini, its largest AI model, to take on OpenAI
"Gemini Ultra was built to quickly understand and act on different types of information — like text, images, audio, video and code. With its multimodal reasoning capabilities, Gemini Ultra can understand, explain and generate high-quality code in the world's most popular programming languages," Hsiao said.
Bard, which was first unveiled by Google in February 2023 as an answer to OpenAI's ChatGPT, was previously running on the company's PaLM 2 large language model.
Story continues below Advertisement
Hsiao said that Gemini Pro outperformed OpenAI's GPT-3.5 in six of eight benchmarks including in MMLU (massive multitask language understanding) and GSM8K (Grade School Math 8K), which measures grade school math reasoning, before its public launch.
MMLU uses a combination of 57 subjects such as math, physics, history, law, medicine and ethics for testing both world knowledge and problem-solving abilities while GSM8K comprises 8,500 high quality grade school math problems that require multiple step reasoning.
OpenAI recently released GPT-4 Turbo, an improved version of the AI upstart's flagship GPT-4 model. However, the free version of its chatbot ChatGPT is still powered by GPT-3.5. To access newer OpenAI models, users must sign up for a monthly Plus subscription.
It will be interesting to see whether Google takes a leaf out of OpenAI's playbook and make Bard Advanced a subscription-based product, offering a cutting-edge AI experience to consumers.
In a briefing with reporters, Hsiao said they didn't have anything specific to share on this matter. "Right now, our focus is delivering the best product experience for people with Gemini. As the experience develops, we will explore how monetisation may look like" she said.
Hsiao said that they are currently completing trust and safety checks and will launch a "trusted tester" programme for Bard Advanced soon with a broader rollout early next year.
In recent months, Google has added several new capabilities to Bard, amid a broader push across the company to integrate generative AI capabilities across their suite of products.
Bard now has deep integration with several of the tech giant's consumer applications and services, including Gmail, YouTube, Docs, and Maps. It can also now understand YouTube videos and answer specific questions regarding the content of some of these videos.
In addition, the firm expanded Bard's visual search capabilities, including the ability to upload images with Google Lens and get Google Search images in responses, along with the chatbot's ability to change the tone and style of responses to more than 40 supported languages and countries.
In October 2023, Google debuted Assistant with Bard, a virtual assistant that combines Bard's generative and reasoning capabilities with Assistant's personalised help.
"Bard's language understanding, complex reasoning and generative capabilities have fueled our excitement in realizing our vision for Google Assistant. It's a step towards our vision to deliver the world's most helpful personal assistant," Hsiao said at the time of the announcement.
Google's AI chatbot Bard gets its biggest upgrade with Gemini launch - Moneycontrol
Read More
No comments:
Post a Comment