Had so much fun on stage at #GoogleIO today – lots of Gemini magic! ✨. ICYMI, Gemini 2.5 is just getting better and better! .So proud of the whole team, and this is still just the beginning. Highlights below 🧵

RT @sundarpichai: New from our security teams: Our AI agent Big Sleep helped us detect and foil an imminent exploit. We believe this is a f….

RT @OfficialLoganK: Today we are rolling out our first Gemini Embedding model, which ranks #1 on the MTEB leaderboard, as a generally avail….

We're imagining a future where Gemini can see what you see -- as @AniBaddepudi says, "Everything is vision". Catch Ani & @OfficialLoganK talking about Gemini's SOTA ability to understand videos, images, documents, how we got here and where we're going!

Welcome to the team, @_mohansolo, Douglas, and crew! Excited to make magical dev experiences together.

If you’re prepping for the IIT JEE, give Gemini a try — it’s pretty good 🧠.

Have you tried Gemini CLI yet!!.

Whatever you need -- Gemini has got you :)

2.0 Flash Lite Preview is our fastest model yet, and improves on quality over 2.0 Flash Lite across the board. You can complete complex tasks in seconds, and with it's cost, it's also great for scale.

gemini-2.5-pro .gemini-2.5-flash. No previews & dates :) . After iterating with all of you, we've graduated our best models to be stable, production-ready, models! These models will have long-term support. And, we're adding our speedy Flash Lite to the family!

Take a look at these metrics! .Gemini 2.5 Pro leads in areas like AIDER, an editing benchmark. It continues to be awesome for code and reasoning

Gemini 2.5 Pro is getting better based on your feedback, and we hope to make this version our stable release soon!. The new 2.5 Pro Preview improves benchmarks and:.1) Introduces budgets for more control.2) Is a more well-rounded model with better creativity, style, & formatting

And. the models are powering even more. Today at I/O, you saw the impact of Gemini behind amazing new experiences in Gemini App, Search, Glasses!!. Read about all the Gemini updates from #GoogleIO:

We're partnering with companies like @AutomationAnywh, @UiPath, @browserbasehq, @autotabai, @interaction and @getcartwheel to test and iterate on bringing Project Mariner's computer use capabilities to the Gemini & Vertex APIs soon!.

Gemini 2.5 is also available in all your favorite IDEs! .On @cursor_ai -- 2.5 Pro is the fastest growing model of 2025 🚀🚀🚀. We also introduced our asynchronous coding agent, Jules -- go to to try it out.

We're making Gemini more usable for developers, with thought summaries in the API, thinking budgets coming to 2.5 Pro, better security, and more natural TTS!. Also, the Live API now has native audio dialogue, which means you can build much more natural, conversational experiences.

2.5 Pro DeepThink brings our research in thinking research to 2.5 Pro, in a mode where the model explores multiple hypotheses before responding. It is SOTA on math, code, & multimodal benchmarks! . We're partnering with safety testers to get this model right -- more soon!

The new 2.5 Flash is unreal quality for cost, with better performance AND 20-30% more token efficiency on our evals

Today's update of Gemini 2.5 Pro Preview is amazing at code web dev, editing, and transformation -- it is now #1 on the WebDev arena, with an ELO score of 1420. That means you can build super cool, visually stunning web apps, with just a single prompt! . Check out this example,

We're 14 days until I/O, and we just dropped Gemini 2.5 Pro (I/O edition)! It's an amazing coding model, and we're so excited to see what you can build with it. We want to highlight the best demos and use cases you build with this model at I/O, so head to