So excited to share our work on the new frontier for generally capable agents in 3D worlds!! Some of the most exciting showcases of agentic capabilities in the blog post, including on entire unseen Genie generated worlds 😜

We should all write papers like this:

I'm so happy to share that I’ll be joining @UofT as an Assistant Professor of Statistical Sciences and Computer Science, with an appointment at the @VectorInst, in 2026! I'm recruiting postdocs and PhD students: https://t.co/CMLLqktRe1! Please help me spread the word! 🧵(1/5)

We just dropped Nano Banana Pro, built on Gemini 3. 🍌 With state-of-the-art text rendering, vast world knowledge and studio-quality creative controls, Gemini 3 Pro Image can create and edit more complex visuals, infographics and more. Here’s what’s under the hood. 🧵

The secret behind Gemini 3? Simple: Improving pre-training & post-training 🤯 Pre-training: Contra the popular belief that scaling is over—which we discussed in our NeurIPS '25 talk with @ilyasut and @quocleix—the team delivered a drastic jump. The delta between 2.5 and 3.0 is

For example I’ve been doing a bunch of late night vibe coding with Gemini 3 in @GoogleAIStudio, and it’s so much fun! I recreated a testbed of my game Theme Park 🎢 that I programmed in the 90s in a matter of hours, down to letting players adjust the amount of salt on the chips!

This is Gemini 3: our most intelligent model that helps you learn, build and plan anything. It comes with state-of-the-art reasoning capabilities, world-leading multimodal understanding, and enables new agentic coding experiences. 🧵

@code_star I heard Gemini 3 answers questions before you ask them. And that it can talk to your cat.

Announcing SIMA 2, the next step in generalist agents for 3D virtual worlds. A significant boost in embodied performance, generalization to new environments, and self-improvement. Looking forward to talking about it more at NeurIPS:

this is absolutely wild... SIMA 2: an agent that can play, reason, and learn inside any 3D game world one of SIMA's 2 most exciting new capabilities is its capacity for self-improvement. from the DeepMind research: "we've observed that, during training, SIMA 2 agents can

Super excited to announce SIMA 2! It’s a general agent that can understand & reason about complex instructions and complete tasks in simulated game worlds, even ones it has never seen before. Incredible to see how it can learn just from self-play… a crucial step towards AGI

SIMA 2 is such a cool and exciting project! An AI agent that interacts in 3D virtual worlds, that you can interact with, talk to, and play games alongside it. We have such an incredible portfolio of amazing projects @GoogleDeepMind !

Very excited to share our work on SIMA, a general embodied agent for 3D worlds. Even more excited to share our work on self-improvement where our gemini-based SIMA 2 agent self-improves to human performance (no humans needed) using a gemini utility model and a gemini task

Introducing Sima 2, our self-improving gemini agent in 3D worlds. 🤖 We trained gemini models to act as embodied agents that perceive, understand, and takes actions in complex 3D worlds. Sima 2 can self-improve in entirely new environments. 🚀 https://t.co/62nxGmBEUb

Training generalist agents on infinite worlds!!!! ♾️♾️♾️

Introducing SIMA 2, the next milestone in @GoogleDeepMind’s research creating general and helpful AI agents. 🤖 Powered by Gemini, SIMA 2 can follow human-language instructions in virtual worlds and also think about its goals, converse with users and improve itself over time.

Google DeepMind's SIMA 1 vs SIMA 2 The bitter lesson continues to be bitter sweet

this is a much, much better path to AGI than any current attempt you need to game max and randomize max and simulation max

So excited to share SIMA 2! We've made incredible progress in creating generalist, interactive, embodied agents in complex 3D worlds. And SIMA 2's capacity for self-improvement is a glimpse of what's to come!

So excited for the SIMA agents work ! This is yet another one where 1+1=3 for @GoogleDeepMind SIMA agents and Robotics 🦾 🧠 = 🤯

3 years ago I started the SIMA project with the dream of using 3D games as worlds in which to train and test Gemini based game agents — all as a stepping stone towards real world AGI. See the linked DeepMind blog post for more information about what the team has since achieved!