Although generative AI is ubiquitous, the majority of implementations rely on the GPT model, which also drives Microsoft Copilot and ChatGPT. Google has been developing an alternative based on its LaMDA framework, and it has now acquired Gemini, which is rumored to be more potent.
Google Gemini AI is an AI model that, according to the company, is more expansive and proficient. Google stated in a blog post, “Gemini is the outcome of extensive collaborative endeavors involving teams from various departments within Google, including our research colleagues.” “It was designed from the foundation up to be multimodal,” which means it can comprehend, operate across, and combine various forms of data (text, code, audio, image, and video) in a seamless manner.
Possibly the most distinctive aspect of Gemini is its variety of editions. Gemini Ultra is designed to operate on robust server hardware and is the most capable version; on the other hand, Gemini Pro is the “best model for scaling across a broad spectrum of tasks.” Ultimately, Gemini Nano possesses the capability to operate locally on devices equipped with specialized hardware. The Recorder app on the Pixel 8 Pro is already utilizing Gemini Nano to summarize content wholly on-device and without an internet connection. In Android 14, app developers can utilize the AICore service to implement custom AI tasks.
According to internal Google reports, Gemini Ultra exhibits marginally higher performance than GPT-4 in multi-step reasoning, reading comprehension, fundamental arithmetic, and Python code generation benchmarks. The comparative performance of Gemini and the more recent GPT-4 Turbo model, which was introduced to ChatGPT Plus subscribers in November and is presently undergoing testing by Microsoft in Copilot (previously known as Bing Chat), remains uncertain.
Beginning today, Google’s Bard AI chatbot will utilize “a fine-tuned version” of Gemini Pro to generate more complex queries. Google describes this as the most significant update to Bard since its inception. The Pixel 8 Pro is equipped with Gemini Nano, and Google is conducting experiments to determine whether or not updating its AI web search with Gemini could improve quality and reduce latency. A future version of Gemini will also be available in Google Chrome, although the functionality and purpose of this version remain unknown.