Skip to content

Updated May 23, 2026 · Reflects Google I/O 2026

beginner path

Gemini & Google AI Mastery

Master Google's full AI stack — Gemini 3.5 Flash, Gemini 3.1 Pro, Antigravity 2.0, NotebookLM, Veo 3.1, and Nano Banana Pro — from your first prompt to production agents.

12 hours
8 videos

What You'll Learn

  • Pick the right Gemini model — 3.5 Flash, 3.1 Pro, Deep Think — for each task
  • Build production agents with Antigravity 2.0 (Desktop, CLI, and SDK)
  • Turn any source set into briefings, audio, and cinematic video with NotebookLM
  • Generate 2K/4K images with Nano Banana 2 (Flash) and Nano Banana Pro
  • Produce film-grade video with Veo 3.1 and Gemini Omni
  • Ship multimodal apps via Google AI Studio and Vertex AI
  • Optimize cost vs. quality across the full Google AI stack

Google shipped its biggest developer release in years at I/O 2026 (May 19): Antigravity 2.0 as a standalone agentic platform, Gemini 3.5 Flash going GA as the new default, Gemini Omni unifying multimodal generation, and NotebookLM gaining cinematic Video Overviews. This portal pulls the official launches and the sharpest expert walkthroughs into one immersive path — so you can move from your first AI Studio prompt to shipping production agents without sifting the noise.

Think of it as a map: Gemini 3.1 Pro and 3.5 Flash are the brains, Antigravity is where agents live, NotebookLM is the research and briefing surface, and Veo 3.1 + Nano Banana Pro handle the media. Each tool is linked below with its official Google page; each video below is curated for signal, not view count.

Start with Foundations (videos 1–4) for the model layer, then branch into the track that fits — Builder (Antigravity, CLI), Researcher (NotebookLM), or Creator (Veo, Nano Banana). The Ecosystem grid, I/O 2026 timeline, and FAQ give you the lay of the land in one screen.

Course Videos

A new era of intelligence with Gemini 3
beginner
See YouTube

A new era of intelligence with Gemini 3

Google DeepMind's official launch film for Gemini 3 — the foundation behind every model variant in this portal.

by Google DeepMind
geminilaunchfoundations
Introducing Gemini Omni
beginner
See YouTube

Introducing Gemini Omni

Logan Kilpatrick walks through Gemini Omni — the new "anything from any input" model unveiled at I/O 2026.

by Google DeepMind
gemini-omnimultimodalvideo
How Pros Use Gemini 3.0 with Google DeepMind's Logan Kilpatrick
intermediate
See YouTube

How Pros Use Gemini 3.0 with Google DeepMind's Logan Kilpatrick

Deep-dive on practical Gemini 3 Pro workflows — vibe coding in AI Studio, agentic patterns, and where 3.1 Pro shines.

by The Neuron
gemini-3-proworkflowai-studio
Gemini 3.1 Pro, Deep Think, New AlphaFold... Buckle Up!
intermediate
See YouTube

Gemini 3.1 Pro, Deep Think, New AlphaFold... Buckle Up!

Benchmark-driven analysis of Gemini 3.1 Pro's Deep Think mode and what HIGH thinking level actually unlocks.

by AI Explained
deep-thinkreasoningbenchmarks
Google's NEW Antigravity 2.0 Is Here — Full Walkthrough for Beginners
beginner
See YouTube

Google's NEW Antigravity 2.0 Is Here — Full Walkthrough for Beginners

End-to-end walkthrough of Antigravity 2.0: install, set up the agent harness, and ship your first website.

by Indie Builder
antigravityagentsbuilder
Google Antigravity CLI Full Walkthrough
intermediate
See YouTube

Google Antigravity CLI Full Walkthrough

Terminal-first tour of the Antigravity CLI — the successor to Gemini CLI, sharing the same agent harness as the desktop app.

by Indie Builder
antigravity-cliterminalagents
NotebookLM Changed Completely: Here's What Matters (in 2026)
beginner
See YouTube

NotebookLM Changed Completely: Here's What Matters (in 2026)

Jeff Su's 2026 NotebookLM tour — the new Studio panel, Reports, Interactive Audio Overviews, and Video Overviews.

by Jeff Su
notebooklmresearchworkflow
15 Ways to Use Nano Banana Pro: Google's Best AI Image Model
intermediate
See YouTube

15 Ways to Use Nano Banana Pro: Google's Best AI Image Model

Fifteen production-grade workflows for Nano Banana Pro (Gemini 3 Pro Image): 2K text rendering, character consistency, and scene composition.

by Creator Tutorial
nano-bananaimage-gencreator

The Ecosystem

Every tool you need to ship — with the official Google product page for each.

ReasoningUpdated

Gemini 3.1 Pro

Google's flagship reasoning model. Set thinking level HIGH for Deep Think mode on complex code, research, and multi-step planning.

SpeedNew

Gemini 3.5 Flash

GA since I/O 2026. Default model across Antigravity. Outperforms 3.1 Pro on coding and agentic benchmarks at ~4× the speed.

MultimodalNew

Gemini Omni

New "anything from any input" model — starts with video. Rolling out to AI Plus, Pro, and Ultra subscribers via the Gemini app and Google Flow.

AgentsNew

Antigravity 2.0

Standalone agent-first desktop app + CLI + SDK. Replaces Gemini CLI (sunset June 18, 2026). Default model: Gemini 3.5 Flash.

ResearchUpdated

NotebookLM

Audio Overviews with Interactive Mode (raise hand to interrupt the hosts). Cinematic Video Overviews in 10 visual styles. Multi-output Studio panel.

VideoUpdated

Veo 3.1 + Veo 3.1 Lite

Text-to-video up to 60 seconds at 4K with native audio. Lite ships at under 50% the cost of Fast. Available via VideoFX, Vertex AI, and Gemini API.

ImageNew

Nano Banana Pro / Nano Banana 2

Pro = Gemini 3 Pro Image (deliberate thinking pass, 2K native, up to 4K). 2 = Gemini 3.1 Flash Image (Flash speed). 5-subject character consistency, image search grounding.

Builder SurfaceUpdated

Google AI Studio

Fastest way to try Gemini and ship apps. New native Android app launched at I/O 2026 alongside the Gemini API "Get code" flow.

What just shipped

Direct from Google — launches, updates, and deprecations to know about.

  1. 2026-05-19Launch

    Antigravity 2.0 launches as a standalone agentic platform

    Standalone desktop app, CLI, SDK, Managed Agents in Gemini API, and enterprise support. The Antigravity CLI replaces Gemini CLI.

    Read on Google
  2. 2026-05-19Launch

    Gemini 3.5 Flash hits General Availability

    Default model across Antigravity. Outperforms Gemini 3.1 Pro on coding and agentic benchmarks while running about 4× faster.

    Read on Google
  3. 2026-05-19Launch

    Gemini Omni & Gemini Omni Flash launch

    New "anything from any input" model starting with video. Rolling out to AI Plus, Pro, and Ultra subscribers via the Gemini app and Google Flow.

    Read on Google
  4. 2026-05-19Research

    Gemini for Science and new Science Skills in Antigravity

    Co-Scientist, AlphaEvolve, Empirical Research Assistance, and NotebookLM are now bundled for researchers. Register at labs.google/science.

    Read on Google
  5. 2026-06-01Update

    Gemini 3.5 Pro arrives next month (preview)

    Sundar Pichai's I/O keynote confirmed Gemini 3.5 Pro is being used internally and will roll out publicly in June 2026.

    Read on Google
  6. 2026-06-18Deprecation

    Gemini CLI sunset — migrate to Antigravity CLI

    Gemini CLI and Gemini Code Assist IDE extensions stop serving requests for AI Pro, Ultra, and Code Assist for individuals. Antigravity CLI is the successor.

    Read on Google

Who to follow

Official Google channels and the sharpest independent voices in the Gemini ecosystem.

Google DeepMind

Official

Official channel — model launches and research

First-party launch films and research deep-dives. Highest signal source for model behavior.

Visit channel

Google for Developers

Official

Official channel — Gemini API, AI Studio, Antigravity

Developer keynotes, API walkthroughs, and Antigravity tutorials direct from the product teams.

Visit channel

Logan Kilpatrick

Google DeepMind — leads Gemini API & AI Studio

Best signal on roadmap timing and developer experience. Hosts the Gemini Omni launch video in this portal.

Visit channel

Jeff Su

Productivity creator

Sharpest NotebookLM workflow content in 2026 — covers Studio, Reports, and Video Overviews better than anyone.

Visit channel

Matt Wolfe (Future Tools)

AI tool review

Holistic AI tool reviews that span the full Google stack — useful for spotting workflows you missed.

Visit channel

AI Explained

Independent analysis

Benchmark-driven comparisons of Gemini vs. frontier models. Trust this for honest capability calibration.

Visit channel

FAQ

Quick answers for AI search and quick scanning. Each Q&A is also emitted as FAQ structured data.

What is the latest Gemini model in May 2026?

Gemini 3.5 Flash is GA since Google I/O 2026 (May 19) and is now the default across Antigravity. Gemini 3.1 Pro remains the flagship for reasoning, and Gemini 3.5 Pro is rolling out publicly in June 2026.

Is Gemini CLI still supported?

No — Gemini CLI and Gemini Code Assist IDE extensions stop serving requests on June 18, 2026 for AI Pro, Ultra, and Code Assist for individuals. The successor is the Antigravity CLI, which shares the same agent harness as the Antigravity 2.0 desktop app.

What is Antigravity 2.0?

Google's standalone agent-first development platform, announced May 19, 2026. Three surfaces: a desktop app, a CLI, and an SDK. The default model is Gemini 3.5 Flash, and the same agent harness runs across all three so improvements ship everywhere at once.

What is Nano Banana 2 vs Nano Banana Pro?

Nano Banana 2 is the Gemini 3.1 Flash Image model — fast, sub-second generations. Nano Banana Pro is the Gemini 3 Pro Image model — a deliberate "thinking" pass that produces the highest quality output (2K native, up to 4K, character consistency for up to 5 subjects, image search grounding).

How do I use Deep Think mode in Gemini 3.1 Pro?

Set the thinking_level parameter to HIGH in the Gemini API (or pick "Deep Think" in the Gemini app). Deep Think is a reasoning mode within 3.1 Pro, not a separate model — use it for research-grade problems where accuracy trumps latency.

What can NotebookLM do that ChatGPT can't?

NotebookLM grounds every answer in the sources you uploaded — no hallucinated citations. The 2026 release adds Interactive Audio Overviews (raise your hand to interrupt the AI hosts), cinematic Video Overviews in 10 visual styles, and a Studio panel that stores multiple outputs of the same type per notebook.

What's the difference between Veo 3.1 and Gemini Omni?

Veo 3.1 is the dedicated text-to-video model — up to 60 seconds at 4K with native audio. Gemini Omni is the broader "anything from any input" successor that handles video alongside multimodal editing across formats. For pure video generation, Veo 3.1 (and the cheaper Veo 3.1 Lite) are still the sharpest tools.

Which Gemini model should I use for coding agents?

Default to Gemini 3.5 Flash (the new Antigravity default) for speed and cost. Switch to Gemini 3.1 Pro with thinking level HIGH when you need architectural reasoning over large codebases or multi-step plans that benefit from Deep Think.

How does Gemini compare to Claude for development?

Gemini wins on multimodal, 1M+ token context, and native agent infrastructure (Antigravity). Claude leads on long-form refactoring and writing. See our Frontier Model Landscape 2026 for the head-to-head with code samples.

Where do I start if I'm new to Google AI?

Watch video 1 (Gemini 3 launch) and video 5 (Antigravity 2.0 walkthrough) in this portal. Then create a free key at aistudio.google.com, try a multimodal prompt, and follow the Builder track from there.

Related Resources

Related Article
Related Article
Related Article
Related Article
Guide

Ready to ship with Gemini?

Pair this portal with our hands-on guides and architecture playbooks. Build an agent, ship a multimodal app, or deep-dive how Gemini compares to Claude.

Explore GuidesMore Learning Paths