Google is attempting to create waves with Gemini, a new generative AI platform that recently made its public debut. However, while Gemini appears to be promising in some ways, it falls short in others. So, what is Gemini? How may it be used? And how does it compare to the competition?

To help you stay up to date on the newest Gemini advancements, we've created this handy guide, which will be updated as new Gemini models and features are launched.

What is Gemini?

Gemini Ultra, the flagship Gemini model; Gemini Pro, a "lite" Gemini model; and Gemini Nano, a smaller "distilled" model that works on mobile devices such as the Pixel 8 Pro.

All Gemini models were taught to be "natively multimodal" — that is, capable of working with and using more than simply text. They were pre-trained and fine-tuned on a wide range of audio, photos, and videos, as well as a big number of codebases and texts in many languages.

What can Gemini do?

Because Gemini models are multimodal, they may theoretically do a variety of tasks, including speech transcription, image and video captioning, and artwork generation. 

Google badly underperformed with the initial Bard launch. More recently, it stirred feathers with a film professing to demonstrate Gemini's capabilities, which turned out to be extensively doctored and more or less aspirational.

Gemini Ultra

So yet, only a "select set" of consumers from a handful of Google products and services have had access to Gemini Ultra, the "foundation" model upon which the rest are constructed. That won't change until later this year, when Google's largest model is released more freely. Most of the information about Ultra comes from Google-led product demos, so take it with a grain of salt.

According to Google, Gemini Ultra can help with physics homework by answering questions step by step on a worksheet and pointing out probable flaws in previously completed solutions.

Gemini Pro

Gemini Pro, unlike Gemini Ultra, is now widely available.

However, its capabilities are unclear because they vary depending on where it is employed.

Google claims that in Bard, where Gemini Pro was first released in text-only format, the model outperforms LaMDA in terms of thinking, planning, and comprehending. A separate investigation by Carnegie Mellon and BerriAI researchers discovered that Gemini Pro outperforms OpenAI's GPT-3.5 in handling longer and more complicated reasoning chains.

Is Gemini better than OpenAI's GPT-4?

There's no way to tell how the Gemini family stacks up until Google launches Ultra later this year, but the company has claimed improvements over the current state of the art, which is often OpenAI's GPT4.

Google has repeatedly emphasized Gemini's advantage in benchmarking, claiming that Gemini Ultra outperforms current state-of-the-art results on "30 of the 32 widely used academic benchmarks used in large language model research and development." According to the business, Gemini Pro outperforms GPT-3.5 in tasks such as content summarization, ideation, and writing.