A few weeks after the rest of the world, the European Union is finally entitled to Gemini Pro, the new multimodal language model that Google presents as equivalent or superior to GPT-3.5. It is possible to try it for free from Google Bard.
Good news for European Google Bard users: the chatbot is becoming more intelligent.
After launching Gemini, its new language model, in English-speaking countries in December, Google is extending its new service to all languages on February 1. Gemini Pro, the intermediate model (There is also Gemini Nano for local tasks and Gemini Ultra for more advanced work), is now the language model used by Bard to answer questions from European Internet users. PaLM2, its predecessor, is bowing out.
Google vs. OpenAI
Little by little, Google is catching up on OpenAI. The web giant, which had probably not anticipated the advance of ChatGPT when it was released in November 2022, talked about generative artificial intelligence throughout 2023… without convincing. Despite its history and know-how in this area, Google has long seemed too far away, particularly when it launched Bard, an intelligent assistant much less relevant than ChatGPT.
At the end of 2023, with Gemini, Google reassured users and investors. Gemini is a language model that addresses both local and very complex tasks, with the ability to process text, image and sound. The first filmed demonstrations of Gemini Ultra, although shortened, were extremely encouraging. The company finally seems poised to catch up with OpenAI.
Between the announcement of Gemini and the deployment of Gemini Pro in France, Google deployed Gemini Nano, its local language model, on the Pixel and Galaxy S24 smartphones. Thanks to it, these smartphones can carry out tasks without the Internet, such as rewriting a text or summarizing a conversation.
Google takes on Midjourney
Alongside the expansion of Gemini Pro, Google announces that it is improving its image generation model, Imagen 2, which is now capable of generating realistic photos. Imagen 2 can be accessed directly from Google Bard, but only in English at launch.
What next steps for Google? The I/O conference in May should be largely dedicated to artificial intelligence. Soon, Google should also integrate Gemini into Google Assistant. The voice assistant present on Android and Nest speakers could merge with Google Bard. Amazon and Apple, its two main competitors, should do the same thing in 2024 (with Alexa and Siri). Finally, the launch of Gemini Ultra, presented as capable of feats, is also eagerly awaited. It should be a paid version of Bard, like Microsoft Copilot Pro and ChatGPT Plus.
If you liked this article, you will like the following: don’t miss them by subscribing to Numerama on Google News.