Google Announces Bard Advanced, Most Advanced Gemini-Powered AI Helper
Google announced huge updates, including a packed Pixel Feature Drop and its new multimodal AI architecture called Gemini. The latter has three versions: Gemini Nano, Gemini Pro, and Gemini Ultra. All three represent Google’s next-generation AI model but are designed for different sizes and tasks.
If you want to try out Gemini Nano, you can already do it on the Pixel 8 Pro with the latest December update. Similarly, the more advanced Gemini Pro model is available to try out with Google Bard. But Gemini Ultra, Google’s most powerful AI model designed for highly complex tasks, will only come to Google Bard next year. With its arrival, Bard will become “Bard Advanced,” giving users access to an AI that can quickly understand and act on different types of information, including text, images, audio, video, and code.
Google says Bard Advanced will also be one of the first ways people can try out Gemini Ultra. The company is currently completing “extensive safety checks” and plans to launch a “trusted tester program” soon before opening up Bard Advanced to more users in early 2024.
“This aligns with the bold and responsible approach we’ve taken since Bard launched. We’ve built safety into Bard based on our AI Principles, including adding contextual help, like Bard’s Google it button to more easily double-check its answers. And as we continue to fine-tune Bard, your feedback will help us improve,” Google wrote in a blog post.
Google says Gemini Pro outperforms Open AI’s GPT 3.5 model in six out of eight benchmarks, including Massive Multitask Language Understanding (MMLU), one of the essential standards for measuring AI models. It even performs better than GPT in grade school math reasoning.
So with Gemini Pro, Bard should now be more capable at tasks such as understanding, summarizing, reasoning, coding, and planning.
“In blind evaluations with our third-party raters, Bard is now the most preferred free chatbot compared to leading alternatives,” claims Google.
You can now try out Gemini Pro in Google Bard for text-based prompts. Google says support for other modalities will be added soon. The experience will be available in English in more than 170 countries to start and come to more languages and regions, like Europe, in the near future.
Google got YouTuber Mark Rober to test Bard with Gemini Pro to craft a paper aeroplane. You can check out the video above.
