What Is Google Gemini (Formerly Bard)? Everything You Need to Know

What is Google Gemini (formerly Bard)? Everything you need to know about AI

Google Gemini is a multimodal AI system: it understands text, images, audio, video and documents; generates text, images, videos and music; and can carry out multi-step tasks independently. Its former name, Bard, refers to the English word for ‘poet’ or ‘bard’. It is based on a proprietary family of large language models developed by Google DeepMind. Gemini 3.8 Flash (since 2 September 2026) is Google’s latest model, whilst Gemini 3.1 Pro remains the most powerful model in the Gemini app.

The model's top functionalities include:

• Creating images, videos and music: images with Nano Banana, videos with Gemini Omni Flash and Veo, music with Lyria. Videos created with Gemini Omni bear Google’s invisible SynthID watermark.
• Deep Research: According to Google, Gemini searches hundreds of sources and compiles a structured report from them in a matter of minutes.
• Gemini Live: Real-time voice calls, with the option to use a camera or screen sharing.
• Canvas and Gems: Canvas is a workspace for text and code; Gems are custom, reusable Gemini assistants.
• Gemini Notebook (formerly NotebookLM): Research based on your own sources, with audio and video summaries. Notebooks synchronise between the app and Gemini Notebook.
• Integration with Google services: Gemini can access Gmail, Docs, Drive, Calendar or Maps on request, thereby providing more personalised answers.
• Agent-based functions: With Gemini Spark, Google offers an AI agent that carries out tasks round the clock.

Gemini 3 and 3.x: Google’s 2026 model range

Gemini 3 helps you learn, create and plan anything.

Sundar Pichai, CEO Google & Alphabet

Google has significantly accelerated its update cycle in 2026. Gemini 1 was released at the end of 2023, Gemini 2 at the end of 2024, and Gemini 3 Pro on 18 November 2025. Since then, new versions have been released at intervals of just a few weeks: between July and September 2026 alone, Google released three new Flash models. According to Sundar Pichai, Gemini 3 recognises the context and intent behind queries better than earlier models, meaning less prompting is required. Here is an overview of the models:

All Gemini updates can be followed on the Google blog.

How Does Google Gemini (Formerly Bard) Work?

The Original Goal of Google Gemini

The original aim of launching Google’s chatbot was to complement the Knowledge Graph cards that appear in response to simple questions. Whilst Knowledge Graph cards can provide definitions of terms or overviews of people or places, Bard’s (now Gemini’s) responses were intended for what Google calls ‘NORA’ questions: questions without a single correct answer. The launch of ChatGPT in late 2022 certainly also accelerated the rapid roll-out of Bard in March 2023.

How Gemini Works

Google Gemini is based on a Large Language Model (LLM) that has been developed from the ground up to be multimodal. It analyses a query along with its context and uses this information to calculate the most likely appropriate response, step by step.

Since Gemini 3, Google has been using a sparse mixture-of-experts architecture: only part of the model is activated per query, which saves on computing power. The models ‘think’ before providing a response. Developers can control the depth of this reasoning via the thinking_level parameter. For up-to-date information, Gemini draws on Google Search. If you allow it, it also incorporates data from Gmail, Calendar, Drive, Keep or Maps to provide more personalised responses. The aim is to provide a response that fits the context, but it remains a statistical prediction and may be incorrect.

The launch of Gemini in December 2023

Particularly when compared to its rival ChatGPT, the language model Bard – now known as ‘Gemini’ – which was unveiled in December 2023, represents a major step forward. What set Gemini apart from the outset was its architecture: Google trained the model to be multimodal from the ground up, rather than simply adding image or audio comprehension capabilities at a later stage. Whilst such a model allows Gemini (formerly Google Bard) to stand out from competing AIs such as ChatGPT, it is not a unique selling point on the market, as there are already other companies with fully multimodal models.

The Renaming of Bard to Gemini (February 2024)

Google launched the chatbot in March 2023 under the name Bard, which at the time was based on the LaMDA language model. Since February 2024, the chatbot, app and model family have all been uniformly named Gemini. The name is the Latin word for ‘twins’. The rebranding was intended to put an end to confusion between the chatbot and the model. Google is continuing along these lines and bringing further AI products under the Gemini brand: NotebookLM became Gemini Notebook in July 2026, whilst Vertex AI became the Gemini Enterprise Agent Platform. Google Assistant is also being gradually replaced by Gemini on Android devices.

If you want to find out more about the model behind the chatbot, you can find detailed information in the article “Google Gemini explained”.

What about using Google Gemini in Germany?

Gemini can be used free of charge in Germany, in German and across all popular apps. According to Google Help, the Gemini web app is available in more than 70 languages and across over 230 countries and territories, including Germany. Gemini understands German queries and responds in German.

Here’s how you can use Gemini:

• In your browser at gemini.google.com
  As an app for Android (Google Play Store) and iOS (App Store)
• As a desktop app for macOS and Windows
• In Google services such as Gmail, Docs, Meet and Google Search (some of which require a subscription)
• For developers and businesses via Google AI Studio, the Gemini API and the Gemini Enterprise Agent Platform

Google Gemini vs. ChatGPT: What Are the Differences?

Gemini and ChatGPT are now capable of similar things, but differ primarily in terms of their ecosystems, models and availability:

‍

Both chatbots generally work on the basis of natural language processing: you ask a question, and the AI generates a human-like response. Both can search the web, analyse files, generate images and carry out multi-step tasks.

The biggest difference lies in the ecosystem. Gemini is built directly into Google’s services, which many people use every day anyway, and can access your emails, calendar and documents with your permission. ChatGPT relies more heavily on its own platform, which can be connected to external services via plugins and MCP. Those who work within the Google ecosystem – using Gmail, Docs and Android – will benefit most from Gemini. Those who prefer to work independently of any particular ecosystem will find ChatGPT offers greater flexibility.

How Can I Use Google Gemini in Germany?

Using the Free Tier Version of Gemini

• Go to the URL gemini.google.com
• No login is required to use Google Gemini; a prompt, i.e. an input, can be sent to Gemini directly in the user interface
• The operation is similar to ChatGPT — after entering the so-called “prompt”, you will receive human-like answers based on your request

According to Google, the free account gives you access to Gemini Flash models, limited access to Pro models, as well as image generation, Deep Research, Gemini Live, Canvas, Gems and Gemini Notebook.

Using the Paid Google AI Pro (Formerly Gemini Advanced) Version of Google Gemini

Google offers three subscription plans for enhanced usage and new features: Google AI Plus, Google AI Pro and Google AI Ultra. The name ‘Gemini Advanced’ is no longer used. Paid subscriptions are available to users aged 18 and over. Google calculates usage limits within the app based on computational effort: the limit resets every 5 hours up to a weekly limit and can be extended using AI credits. Prices vary by region, particularly between US dollar and euro prices. There are also occasional discount offers, such as lower introductory prices for Plus or Pro for students.

Official prices and descriptions: Google Gemini subscriptions.

How Does Logging in to Google Gemini Work?

To sign in to Google Gemini, simply follow these steps:

1. Open the Google Gemini website in your browser (gemini.google.com).
2. Click the “Sign In” button at the bottom left or at the top right.
3. Enter the email address or phone number linked to your Google Account.

Is Google Gemini (Formerly Bard) Free?

Yes, Google Gemini is free to use. All you need is a Google account. According to Google, the free version includes Gemini 3.6 Flash, limited access to Gemini 3.1 Pro, Deep Research, Gemini Live and image generation, although with usage limits. For higher limits, the latest model, video generation and more storage, you’ll need a paid monthly subscription. Subscriptions renew automatically unless you cancel them in good time.

What Should You Not Share With Google Gemini?

As a general rule, you should not share any personal or confidential information with Google Gemini. The reason for this is that Google collects and stores personal data in order to improve its AI. According to Google, data such as chat content, locations and device information is saved to your Google Account by default in order to personalise the user experience and improve its services.

You can manage this in the settings:

• Disable Gemini activities (labelled ‘Keep Activity’ in the English interface): New chats will then no longer be saved to improve the models. According to Google, it will still retain them temporarily in order to provide the service.
• Use temporary chats: These do not appear in your chat history and are not used for personalisation.
• Check linked apps: Decide whether Gemini is allowed to access Gmail, Drive or Photos.

Different rules apply to businesses: for Google Workspace and Gemini Enterprise, Google undertakes not to use customer data to train its models without permission. In addition, the EU AI Act will impose transparency obligations on providers of large AI models from August 2025. The obligations under the GDPR apply in any case.

Google Gemini as a Helpful Additional Research Tool

Gemini is a powerful tool for research, writing, analysis and creative projects. However, as with any generative AI, answers may be incorrect or incomplete. Be sure to check them, particularly if you are using them to inform decisions or for publication.

An open AI assistant isn’t enough when it comes to communicating with your customers. You need to be able to specify which sources the AI draws on for its responses, which statements are off-limits, and where the data is processed. That’s exactly what moinAI is for: the Agentic AI platform for customer service and marketing, which gives you full control over every response.

[[CTA headline="Try moinAI now and experience the future of customer communication " subline="In just four easy steps,  create a chatbot prototype and get a first impression of the technology — without obligation." placeholder="Insert your website URL here!" button="Try it now!"]]

_{The information in this article is regularly updated to ensure accuracy and truthful content.}