Gemini (translated from English — “twin”) is a chatbot with artificial intelligence, which is based on a family of conversational neural language models developed by Google.
It all started on February 6, 2023. On this day, Google introduced the Bard chatbot to a limited circle of users, which was supposed to answer questions posed in the Google search engine. And already on February 8, 2023, a big presentation of Google Bard took place in Paris, where the service made a mistake answering questions about the latest discoveries made with the help of the James Webb Space Telescope. This unfortunate mistake cost Google $ 100 billion (that’s the amount by which Google shares fell that day).
Over the next few weeks, 80,000 Google employees who responded to management's requests to find and fix Bard's bugs harshly criticized Bard in internal chats for a range of security and ethics issues. access to Bard. However, not wanting to lose the competitive battle with Microsoft and OpenAI, which have teamed up to make a technological breakthrough in the Bing search engine, Google announced preliminary access to Bard.
On March 21, 2023, Bard was given access to select users in the US and UK, who were warned that the chatbot "may provide inaccurate or offensive information that does not align with Google's views." For each question asked of Bard, three answers were given, and users were asked to select the most helpful one.
In April, the developers announced that Bard had gained "mathematical and logical capabilities" and had learned to program: it writes, debugs, and explains program code in 20 programming languages.
In May, Bard became available to the general public in 180 countries, supporting additional languages and integrated into other Google products.
In December 2023, Bard switched to a new language model — Gemini Pro.Bard can now extract information and meaning from images and videos, recognizing and labeling details; "understand" human language, regardless of manner of communication, tone, pitch, accent, etc.; process natural, human-created text, generate articles on the topic, analyze documents, summarize, make predictions and give recommendations; generate an image.
In February 2024, Bard received a new name — Gemini.
As of early 2025, Gemini supports over 40 languages (including Ukrainian): "understands" human language (written and spoken), generates content, writes and speaks; creates realistic images; has full-fledged applications for Android and IOS.
On March 30, 2025, the Gemini 2.5 Pro model was released to the public, which has built-in "thinking" mechanisms for solving complex problems. It has significantly improved coding skills. The model shows good results in a number of mathematical and scientific tests.