Google Gemini just dropped a stack of updates that could seriously change how people build with AI. We are talking about Gemini Flash 3.6, a new lower cost Gemini 3.5 model, a powerful Gemini Omni video model, and major upgrades to what was previously known as NotebookLM.
This is not just another small model refresh. Google is pushing harder into fast agentic workflows, multimodal AI, source grounded data analysis, video creation, conversational video editing, and cross app productivity. If you use Gemini for work, content, automation, research, or building AI tools, there is a lot here to pay attention to.
The big takeaway is simple: Gemini is becoming more capable while also becoming easier to organize, more flexible to use, and in some cases, less expensive to run.
Gemini Flash 3.6: Better Performance at a Lower Output Cost
The first major release is Gemini 3.6 Flash. Flash models have always been about balancing performance and speed, but this release is positioned as a much stronger option for agentic and multimodal tasks.
That matters because modern AI work is not just about asking a chatbot a question. We increasingly need models that can reason through multi step tasks, work with images and video, handle more context, and support automated workflows without becoming painfully slow or expensive.
Gemini 3.6 Flash is designed for exactly that kind of use case.
What Gemini 3.6 Flash Improves
- Stronger agentic performance: Better suited to tasks where AI needs to follow a process, use tools, or complete multiple steps.
- Improved multimodal capability: Built for working with more than plain text, including images, video, and other rich inputs.
- More current knowledge: Its listed knowledge cutoff is March 2026, compared with January 2025 for Gemini 3.5 Flash.
- A larger context capability: More room to work with information in a single interaction.
- Lower output pricing: Listed at $1.50 for input and $7.50 for output, compared with $1.50 input and $9 output for Gemini 3.5 Flash.
That combination is huge. A model that is better, has more up to date knowledge, supports richer tasks, and costs less on output is exactly the direction people want these platforms to move.
For builders, this can make a meaningful difference. When an application needs to generate a lot of output, output pricing is often where costs start adding up. Lower output costs can make it more realistic to use a stronger model for customer support, research tools, content systems, internal AI assistants, and automated workflows.
Gemini 3.5 Lite and the Focus on Efficient Production Work
Google also introduced Gemini 3.5 Lite, described as a fast and cost effective Gemini 3.5 option for high volume production execution. This is the kind of model that makes sense when speed, reliability, and scale matter more than pulling out every last bit of reasoning power.
There is also a clear tradeoff to understand: cheaper models can be incredibly useful, but a lower priced model does not automatically mean it is the right one for every job. If you are handling complex analysis, long documents, high stakes decisions, advanced tool use, or rich multimodal tasks, the more capable Flash or Pro options may be the better fit.
The point is not to blindly choose the newest model. The point is to match the model to the job.
A Practical Way to Choose a Gemini Model
- Use Gemini 3.6 Flash when you need a strong mix of speed, intelligence, multimodal capability, and agentic performance.
- Use Gemini 3.5 Lite when you are running high volume, cost sensitive tasks that need fast execution.
- Use Gemini Pro or Thinking models when deeper reasoning or more advanced problem solving is the priority.
- Test the same task across models before committing a workflow to one option.
At launch, the new models appeared in Google AI Studio first, before broader availability inside Gemini. Shortly afterward, Gemini 3.6 Flash and Gemini 3.6 Thinking began appearing across Gemini accounts as the rollout expanded.
NotebookLM Is Becoming Gemini Notebook
The second major update is a rebrand with much bigger implications underneath it. NotebookLM is now being called Gemini Notebook.
On the surface, a name change may not seem like a big deal. Under the hood, though, Google is clearly positioning Notebook as a more connected part of the Gemini ecosystem. The product is moving beyond being a useful research notebook and toward becoming a more capable workspace for organizing source material, analyzing information, and creating outputs based on your own documents.
Collections Make Notebook Organization Far Easier
One of the immediately useful features is Collections. If you have a growing number of notebooks, this is a big quality of life improvement.
Instead of having a long, messy list of unrelated notebooks, you can group them by project, area of work, client, hobby, topic, or workflow. For example, you could create collections for:
- AI agents and automation projects
- Make.com systems and integrations
- Car research and hobby notes
- Content planning and campaign research
- Client deliverables and company knowledge
You can also add emojis to collections. That may sound small, but visual labels make it much faster to scan a workspace when you are juggling lots of projects.
Native Code Execution for Better Source Grounded Analysis
The more important change is that Gemini Notebook is gaining the ability to write and execute code natively. This opens the door to deeper data analysis that stays grounded in the sources you provide.
That is a massive step forward for anyone working with documents, datasets, reports, research materials, or internal knowledge. Rather than simply summarizing information, Gemini Notebook can move toward more complex analytical work and new output formats.
The rollout begins with Google AI Ultra and Workspace users, with broader access for Pro users on the web expected afterward.
This is the kind of feature that can turn a notebook from a place where you store information into a workspace where you actively do something with it. That is a very different level of usefulness.
Cross App Sync Makes Gemini Notebook More Flexible
Another important upgrade is syncing between Gemini Notebook and the Gemini app. Your notebooks can follow you between experiences, so you are not trapped in one isolated tool.
You can access and create notebooks within the Gemini app, with cross app syncing between the app and the Gemini Notebook experience. Google also plans to bring notebooks directly into AI Mode in Search.
That makes the long term direction pretty obvious. Google wants research, search, AI conversations, notebook organization, and data analysis to feel like parts of one connected workflow rather than separate products.
Gemini Omni Flash Preview Changes AI Video Creation
The update that may get the most attention is Gemini Omni Flash Preview, Google’s new video generation and conversational editing model.
This is where things get pretty crazy. Gemini Omni is built to create video from text and images, but it is not limited to generating a video and calling it done. It can also refine results through natural language editing requests.
That means you can create something, inspect it, and then ask for specific changes without having to start the entire process from scratch.
What You Can Control in Gemini Omni
Gemini Omni gives you several creation controls, including:
- System instructions
- Aspect ratio options such as 16:9 and 9:16
- Video duration
- Resolution
- Frame rate
- Thinking level
Those controls matter because AI video is not one size fits all. A vertical short form social video needs a different format from a landscape product demo, presentation asset, or cinematic B roll clip.
From a Simple Prompt to Realistic B Roll
A prompt asking for a King Charles Cavalier running through a Montana field produced a highly realistic video result, complete with changing camera angles, movement, detailed facial features, and natural looking environmental shots.
That kind of output has obvious uses for B roll, social content, concept videos, visual storytelling, ads, and standalone creative projects. The interesting thing is not only the realism. It is that the model can vary shots and camera perspective rather than creating a single static looking scene.
For creators and marketers, that means an AI video tool can begin to feel less like a novelty clip generator and more like a production assistant.
AI Avatars and Conversational Video Editing
Gemini Omni can also create talking AI avatar videos. In one example, an avatar explained Model Context Protocol, or MCP, while moving naturally, turning its head, using hand gestures, and showing emotion.
MCP is a standard designed to connect AI models with data and tools. The topic itself is technical, but the important part here is how the video was created: the motion, presentation, and animation were generated with AI.
Then comes the part that really changes the workflow. After generating the video, a simple instruction asked Gemini Omni to add subtitles while changing nothing else. The model generated an updated version with subtitles while preserving the original animation and content.
You could get even more specific with requests such as subtitle colour, outlines, placement, or other visual adjustments. This turns editing into a conversation.
Instead of manually rebuilding a video for every revision, you can describe what needs to change and let the model handle the revision.
Turning Product Images Into Ads
Gemini Omni is not limited to text prompts. You can upload an image and turn it into a video ad. In a product example, an image of a Rolex Day Date was transformed into a polished luxury style advertisement with voiceover and cinematic presentation.
The result was strong overall, though the day display on the watch needed cleanup. That is a useful reminder: AI video can produce impressive creative assets, but you still need to inspect details carefully, especially in product visuals, text, numbers, logos, and fine design elements.
Still, the workflow is incredibly powerful:
- Start with a product image or concept image.
- Describe the style, pacing, environment, and message.
- Generate a video asset.
- Review the details.
- Request targeted edits in plain language.
That can make creating product ads, explainer clips, social assets, and visual concepts dramatically faster.
How to Get More Out of Google Gemini
New models are great, but the real value comes from using them well. Most people are not getting anywhere close to the full value available from Gemini, ChatGPT, or Claude because they are using generic prompts, skipping automation opportunities, and never reviewing their actual workflows.
A more useful approach is to audit how you use AI for your specific role. Think about the tasks you repeat, the tools you already live in, the work you wish could be automated, and the prompts you keep recreating.
Tools such as LLM Helper are designed to help structure that process. An AI usage audit can identify:
- Prompt templates worth saving and reusing
- Automations that could remove repetitive work
- Google Gems that could support a specific workflow
- MCP servers and connectors worth considering
- Optimization opportunities based on your role and tools
The key is specificity. “How can I use Gemini better?” is too broad. A much better question is, “I work in this role, use these tools daily, repeat these tasks every week, and want to automate these processes. What should I set up?”
Run that kind of audit regularly. AI tools move quickly, and the best workflow today may not be the best workflow a few weeks from now.
The Bottom Line
Google Gemini is moving fast. Gemini 3.6 Flash brings stronger performance with lower output pricing. Gemini 3.5 Lite gives teams another efficient option for high volume work. Gemini Notebook is becoming more organized, more analytical, and more connected. And Gemini Omni is pushing AI video from basic generation into creation plus conversational editing.
The exciting part is not any one feature by itself. It is how these updates can work together. You can research with Gemini Notebook, organize source material into collections, use stronger models for analysis and automation, and create polished video assets from prompts or images.
That is a much more complete AI workflow than simply opening a chatbot and asking a question.
If you are using Google Gemini, this is the moment to revisit your setup. Test the new models. Organize your notebooks. Identify one repetitive task to automate. Try Gemini Omni on a simple video concept. The tools are getting better fast, and the people who build useful workflows around them are going to get the most out of the change.
Share this article with someone building with AI, and explore the latest Gemini tools before your current workflow gets left behind.
Frequently Asked Questions
What is Gemini 3.6 Flash?
Gemini 3.6 Flash is a new Google Gemini model designed to balance speed and intelligence for agentic and multimodal tasks. It is positioned as more capable than Gemini 3.5 Flash while offering a lower listed output cost.
How does Gemini 3.6 Flash compare with Gemini 3.5 Flash?
Gemini 3.6 Flash has a listed input cost of $1.50 and output cost of $7.50, while Gemini 3.5 Flash is listed at $1.50 input and $9 output. Gemini 3.6 Flash also has a March 2026 knowledge cutoff, compared with January 2025 for Gemini 3.5 Flash.
What happened to NotebookLM?
NotebookLM has been renamed Gemini Notebook. The updated experience includes Collections for organizing notebooks, cross app syncing with Gemini, and upcoming capabilities for native code execution and deeper source grounded analysis.
What is Gemini Omni Flash Preview?
Gemini Omni Flash Preview is Google’s video generation and conversational editing model. It can generate video from text prompts and images, then refine generated videos through natural language requests such as adding subtitles or changing visual details.
What can Gemini Omni be used for?
Gemini Omni can be used for B roll, AI avatar videos, product advertisements, image to video creation, standalone clips, and video edits. It supports controls for aspect ratio, duration, resolution, frame rate, and more.
How can I improve my Google Gemini workflow?
Start by identifying repetitive tasks, the tools you use every day, the work you want to automate, and the prompts you frequently recreate. Then build reusable prompt templates, evaluate automation opportunities, and review your AI workflow regularly as Gemini capabilities evolve.
