Although OpenAI’s ChatGPT has surged as a global sensation, quickly becoming one of the most rapidly embraced consumer products, Google’s Bard has lagged behind. Despite acquiring new functionalities like accessing user data from various Google products, Bard’s responses and information have often fallen short compared to ChatGPT and other bots utilizing GPT-3 and GPT-4.
However, the case for Bard might be evolving. As of today, for English-speaking users in 170 countries, Bard now operates on Google’s fresh Gemini model, which Google claims matches or surpasses OpenAI’s technology in multiple aspects. According to Google, Gemini will expand to more languages and regions “soon.”
Bard now utilizes Gemini Pro, positioned as the middle tier among the Gemini models. While Ultra is the largest but slowest and most capable, Nano is compact, swift, and designed for on-device tasks, leaving Pro as the balanced version—swift and efficient yet highly capable.
Sissie Hsiao, overseeing Bard and Assistant at Google, expressed in a briefing that Gemini marks the “most significant and superior upgrade” for Bard. Expectedly, this upgrade should noticeably enhance Bard’s existing functions: summarization, brainstorming, writing, and more. Sundar Pichai, Google’s CEO, noted improved understanding of intent, enhanced factual responses, higher quality output, and better performance, particularly in coding tasks.
Presently, Bard functions solely as a chatbot—text in, text out. However, an upcoming iteration of Bard, called “Bard Advanced,” powered by Gemini Ultra, is slated for next year. Gemini Ultra, the most potent version, is multimodal, capable of handling images, audio, video, and text interactions.
According to Demis Hassabis, head of Google DeepMind, Gemini excels particularly in non-text interactions, showcasing seamless integration across multiple modalities. Google’s demonstrations included scenarios like YouTuber Mark Rober using Bard to perfect a paper airplane, utilizing AI-provided feedback from photos of his designs, and parents seeking help with their children’s homework by uploading images for assistance in understanding math errors.
While these demonstrations are currently limited to promotional materials, Pichai views this launch as a significant milestone for Bard and the inception of the Gemini era. If Google’s assessments hold true, the new model might position Bard on par with ChatGPT, signifying an impressive accomplishment in itself.
