Analyzing Gemini, ChatGPT's new competitor

Google boasts of having a language model superior to ChatGPT and capable of reasoning like a human

A year ago, an artificial intelligence chatbot application developed by OpenAI came to light. Chat GPT uses learning techniques capable of giving responses almost like those a person would give and with a surprising ability to generate texts, computer code or summarize information.

Overnight, GPT Chat became an incredibly popular tool, and an essential part of many users' daily lives, with over 180 million active users.

Gemini is a language model and ChatGPT is a conversational application built on another language model, GPT-4 or GPT-3.5 depending on the version of ChatGPT being considered (paid or free, respectively). In Google's case, ChatGPT's equivalent is Bard , which until now used the PaLM language model but now operates with an adapted version of Gemini for queries in English.

Both Google and OpenAI offer these models on a subscription basis to companies and developers.

Gemini will have three different versions

• Gemini Ultra: Our largest and most powerful model, for highly complex tasks.

• Gemini Pro : Our best model for scaling a wide range of tasks.

• Gemini Nano: Our most efficient model for running tasks directly on a device, available for mobile.

What is Gemini?

Gemini is a multimodal AI model. That is, it is capable of "generalizing and understanding, operating and seamlessly combining different types of information," from text and audio to images and video, including coding languages.

In its presentation, Google notes that the new AI will be the most flexible, largest and most capable it has created to date. Its developers say it is the result of hard collaborative work at scale across multiple teams across its brand, including teams at DeepMind and Google Research.

Gemini will be able to run efficiently on very different devices, from mobile phones to data centers.

Where can Gemini be used?

Gemini 1.0 is now rolling out across a wide range of Google products and platforms, for both consumers and developers.

According to Google's statement, Gemini Pro is available on Bard for more than 170 countries and territories , which significantly improves the chatbot's responses.

Gemini is capable, according to Google, of supporting a context of about 32,000 tokens in questions, which is roughly equivalent to about 32,000 words. That's the same amount that GPT-4 can handle. However, one of Gemini's big advantages will be its ability to reason its answers , giving better results than ChatGPT .

The company also says that one of Gemini's great features will be its ability to code in the most popular programming languages , such as Python, Java, C++ and Go.