Google Gemini’s Crazy New Updates: 3.6 Models, MCP Connectors and Spark Automations

Meta description: Explore Google Gemini 3.6, Chrome skills, Gemini Spark, MCP connectors, Workspace actions, and practical automations for faster AI workflows.

Google Gemini just got a massive set of upgrades, and this is way bigger than a simple model refresh. Gemini 3.6 brings new speed and reasoning options, the Chrome experience can understand the pages and tabs you have open, Gemini Spark can take action across Google Workspace, and MCP connectors make it possible to connect Gemini to a much wider range of tools.

The important shift is this: Gemini is becoming less like a chatbot you ask random questions and more like a personalized assistant that can research, create, organize, act, and improve alongside your existing workflow.

If you are still opening Gemini only to search for an answer or generate a quick paragraph, you are leaving a ridiculous amount of value on the table.

What Is New in Google Gemini?

The biggest changes fall into four buckets:

• New Gemini 3.6 models, including Flash and Thinking options.
• Major Gemini for Chrome upgrades, including page context, multiple tab sharing, screen selection, and reusable skills.
• Gemini Spark improvements, with tasks, schedules, Google Workspace actions, and remote Mac control.
• Custom MCP connectors, allowing Gemini to connect with apps that provide an MCP server.

Each piece is useful on its own. Combined, they unlock some genuinely powerful automations for content creators, operators, researchers, marketers, and anyone trying to get more done with AI.

Gemini 3.6 Models: Flash, Thinking, and What Actually Matters

Gemini now includes two new model options: Gemini 3.6 Flash and Gemini 3.6 Thinking. Gemini 3.6 Pro has not arrived yet, but it is expected to follow.

Inside Gemini itself, this does not necessarily feel like a completely different product. The more interesting upgrade is how these models work inside the Gemini Chrome extension.

The Chrome experience includes several model choices:

• Auto
• Gemini 3.5 Flash-Lite
• Gemini 3.6 Flash
• Gemini 3.1 Pro
• Flash with Extended Thinking

For most everyday work, 3.6 Flash is the exciting one because it is fast while still being smart enough to understand real working context. If you are doing a more complex task, Extended Thinking can be useful. For simple requests, it is probably unnecessary.

The practical takeaway is simple: use the fast model when you need to move quickly across browser tabs, web research, summaries, drafting, and quick analysis. Turn on deeper reasoning when the job actually requires it.

Gemini for Chrome Can Understand What You Are Working On

This is where things start getting crazy. Gemini inside Chrome can reference the page you are currently using, including a YouTube page, article, website, tool interface, or other browser content.

More importantly, Personalized Intelligence can tailor Gemini’s output around what it knows about you. Rather than requesting a generic summary, you can ask Gemini for a summary that focuses only on the parts most relevant to your interests and work.

For example, if you are researching a long video on a topic you care about, you can request a personalized breakdown of the key ideas. Gemini can surface the important takeaways, timestamps, caveats, and even a confidence indicator.

That is a dramatically better workflow than getting a bland list of bullet points and then having to figure out what matters yourself.

Share Multiple Tabs for Better Research

Gemini can also work with multiple shared browser tabs. This is incredibly useful when you have several sources open and need to compare, synthesize, or learn from all of them at once.

Here are a few ways to use multi-tab context:

• Compare five competitor landing pages and identify positioning patterns.
• Review multiple YouTube videos and pull out the most valuable content ideas.
• Analyze several product pages before writing a buying guide.
• Collect research from different articles and turn it into a usable brief.
• Compare software tools, features, pricing language, and customer promises.

Instead of copy-pasting information into a separate chat, the relevant context can already be sitting in the tabs you have open.

Select Anything on the Screen and Ask Gemini About It

The Select from screen feature is another deceptively simple but powerful addition. You can select a specific part of a webpage or interface, capture it, and ask Gemini what it is or what it does.

This can be used to identify someone shown on a page, understand a feature in unfamiliar software, decode an interface element, or get instructions for using a tool.

Imagine opening a new platform, seeing an unfamiliar button or workflow, selecting that area, and asking: “What does this do?” Gemini can provide an explanation based on the visual information you selected.

That makes it much easier to learn new software without constantly digging through help centers, documentation, and search results.

Use Gemini Live When You Get Stuck

Gemini Live has also been upgraded with a newer Live model. This is useful when you are stuck on a process, trying to complete something on your computer, or need help thinking through a task in real time.

The biggest opportunity here is using AI as an active problem-solving companion, not merely a text generator.

Create Reusable Gemini Skills Instead of Rewriting Prompts

One of the best additions is the ability to create skills. A skill is essentially a reusable instruction set that can be called whenever you need it.

Instead of writing the same long prompt every time, create it once, save it, and trigger it using a slash command in a new Gemini chat.

For instance, you might build a skill called “Personalized Video Summary” that tells Gemini to:

• Analyze the current video or page.
• Focus on information related to your work and interests.
• Include timestamps for key moments.
• Flag caveats and uncertainty.
• Deliver concise, actionable takeaways.

Once saved, that workflow is always available. You do not need to rebuild the prompt every time you open a new page.

And summaries are only the beginning. Create skills for the work you repeat all the time:

• Turn research into a YouTube script.
• Transform a video into a blog article.
• Remix an existing piece of content for social posts.
• Create content angles from competitor research.
• Convert meeting notes into action items.
• Generate product comparisons from multiple open tabs.
• Turn an idea into an email, outline, or content brief.

Google has also included prebuilt skill categories around fun, learning, research, shopping, understanding, and writing. But the custom skills are where you can make Gemini fit the way you work.

Gemini Spark Is Becoming an Automation Engine

Gemini Spark is where Gemini starts moving beyond browser assistance and into actual task automation. Spark can help set up skills, create tasks, and coordinate actions across connected apps.

Previously, access was much more limited. Spark is now rolling out more broadly for Pro subscribers, with expanded country and language support across Gemini apps.

If Spark is not appearing on the web despite having a compatible account, it may be worth opening the Gemini desktop app. It can appear there even when it is not immediately visible in the browser experience.

Google Workspace Actions

One of the most useful Spark improvements is its ability to work with Google Workspace. Gemini Spark can help with actions such as:

• Editing private spreadsheets and presentations.
• Reading comments in spreadsheets and presentations.
• Editing shared Google Docs, Sheets, and Slides.
• Adding images to documents and presentations.
• Refining and improving documents.

This matters because valuable work often gets stuck at the handoff stage. An AI can write something, but then someone still has to paste it into the correct document, format it, share it, revise it, and communicate it to the right person.

Spark is designed to help bridge that gap.

Remote Mac Control

Spark also supports remote control for Mac. If your computer is open, you can ask the Gemini app to take actions on your Mac from another device.

That is a huge shift in how personal AI assistants can work. The assistant is no longer confined to generating an answer. It can be positioned to help execute a workflow across devices.

Schedules Turn One-Off Work Into Recurring Systems

Schedules are another major unlock. Rather than asking Gemini to do the same research repeatedly, set it up once and have it run on a recurring basis.

Examples include:

• Finding sponsored emails every day.
• Monitoring homes for sale in a specific city or area.
• Tracking updates in a niche.
• Collecting recurring research for a content pipeline.
• Running regular checks across connected tools.

This is how you start turning AI from something reactive into something proactive.

A Real Gemini Spark Content Automation Example

Here is a great example of what Spark can do. Say you are in the AI news niche and want to create a full YouTube package about ChatGPT Work and ChatGPT 5.6.

You could ask Spark to create:

1. Title ideas.
2. Description ideas.
3. Comma-separated tag ideas.
4. A thumbnail.
5. A full script for the video.

Spark can research the topic, use the relevant skills and connected apps, and show the progress of the task. You can inspect the files it used, the skills involved, and the actions it has taken.

The resulting package can include multiple title angles, such as high-curiosity options, breaking-news angles, productivity and business angles, or search-focused titles. It can create descriptions, tag lists, a thumbnail concept or image, and a detailed script.

The script can go further by suggesting editing direction, screen capture ideas, and B-roll. That is a seriously useful output because the workflow no longer ends at “here is a draft.” It provides a production-ready starting point.

If the thumbnail is not good enough, refine it. Upload a reference screenshot and ask for a better version based on that visual direction. Gemini can generate a revised thumbnail and explain the changes it made to the composition and design.

And if the project is for a client, the workflow can go even further: add the assets to a Google Doc, email that document to the client, and draft the email message. That is the real power of Spark. It connects creation with execution.

Custom MCP Connectors Expand What Gemini Can Access

Custom MCP connectors are one of the most important updates in this entire release. MCP, or Model Context Protocol, makes it possible for Gemini to connect with other applications that have an MCP server available.

In practical terms, this means your Gemini setup can become much more connected to the tools you actually use every day.

Connected apps are already a core part of Spark. The MCP capability expands that idea by making Gemini more flexible across a wider ecosystem of services and workflows.

The goal should not be connecting every tool just because you can. The goal is to connect the systems that hold the information and actions you need most often.

Start with the places where work repeatedly gets fragmented:

• Your documents and spreadsheets.
• Your content research sources.
• Your planning and task systems.
• Your client communication workflows.
• Your recurring reporting and monitoring processes.

Then ask a better question: What information does Gemini need, and what actions should it be able to take, for this workflow to become faster?

Audit Your Gemini Workflow Every Week

Getting access to new Gemini features is not enough. You need to audit how you are using Gemini and identify where it can save time, improve quality, or automate repetitive work.

This is exactly the kind of thing LLM Helper is designed to help with. Its Gemini settings guidance can point you toward the settings that make the most sense for your setup.

It can also help audit your existing Gemini workflow. You provide your role, tasks, and the tools you live in, then use that information to identify better prompts, useful automations, Gemini Gems or skills to create, MCP servers to add, and optimization opportunities.

The point is not to use Gemini like Google Search with a prettier interface. The point is to build an AI system that understands your recurring work and gets better integrated into it over time.

A weekly audit can reveal questions like:

• What prompts am I writing again and again?
• What research tasks can become scheduled tasks?
• What content workflows can become reusable skills?
• Which apps should be connected to reduce copy-pasting?
• What tasks need a better model choice or deeper reasoning?

LLM Helper also includes an LLM starter kit with prompts intended to make ChatGPT, Claude, and Gemini work more like personalized assistants rather than isolated chatbots.

How to Get More Out of Google Gemini Right Now

If you want the practical version, here is the move:

1. Test Gemini 3.6 Flash in Chrome. Use it on your active tabs, research pages, and everyday browser work.
2. Turn on Personalized Intelligence. Ask for outputs tailored to your work instead of generic responses.
3. Create three reusable skills. Start with tasks you repeat every week.
4. Use Select from screen. Make it your default option whenever you are confused by a site or tool.
5. Explore Gemini Spark. Build a task that creates something useful end to end.
6. Create one schedule. Automate a recurring research or monitoring task.
7. Connect the apps that matter. Focus on the tools that contain your key context and workflows.
8. Audit your setup weekly. Keep looking for repetitive work that Gemini can simplify.

Suggested Visual Assets

FAQ

What are the new Google Gemini 3.6 models?

Google Gemini includes Gemini 3.6 Flash and Gemini 3.6 Thinking. Flash is positioned as a fast and capable option, while Extended Thinking can be enabled for more complex tasks. Gemini 3.6 Pro is expected to arrive later.

What can Gemini for Chrome do?

Gemini for Chrome can reference active browser content, use personalized intelligence, work across selected tabs, analyze screenshots through Select from screen, and run reusable skills for common workflows.

What are Gemini skills?

Gemini skills are reusable prompts and workflows that you can save and call when needed. They are useful for repeated tasks such as summarizing content, writing scripts, creating articles, remixing content, and organizing research.

What is Gemini Spark used for?

Gemini Spark can help create skills, run tasks, set schedules, use connected apps, and take Google Workspace actions. It can support workflows involving Docs, Sheets, Slides, content creation, research, image generation, and communication.

What are MCP connectors in Google Gemini?

MCP connectors allow Gemini to connect with applications that offer an MCP server. This can expand Gemini’s ability to access relevant context and participate in workflows across more tools.

Build the System, Not Just the Prompt

These Gemini updates are powerful because they shift the entire experience from isolated prompts to connected workflows. You can bring in browser context, save repeatable skills, run scheduled tasks, work with Google Workspace, and connect more of your stack through MCP.

That is the opportunity: stop treating Gemini as a place to ask one-off questions. Start treating it like an operating layer for the work you do every day.

Pick one repetitive task this week, turn it into a skill or scheduled Spark workflow, and see how much time it saves. Then keep building from there.