Google’s new Gemini 2.5 Pro Experimental takes the #1 position across a range of our evaluations that we have run independently Gemini 2.5 Pro is a reasoning model, it ‘thinks’ before answering questions. Google has released it as an experimental API in AI Studio only, and has

This is indeed really cool, in case you haven't tried it or heard about it.

Gemini 2.5 Pro is taking off 🚀🚀🚀 The team is sprinting, TPUs are running hot, and we want to get our most intelligent model into more people’s hands asap. Which is why we decided to roll out Gemini 2.5 Pro (experimental) to all Gemini users, beginning today. Try it at no

Big update to our MathArena USAMO evaluation: Gemini 2.5 Pro, which was released *the same day* as our benchmark, is the first model to achieve non-trivial amount of points (24.4%). The speed of progress is really mind-blowing.

I've always believed that AGI has the power to revolutionise industries and accelerate science, and that harnessing this power responsibly will be essential. This report explains, in some detail, our current thinking on AGI safety and security.

📣 It’s here: ask Gemini about anything you see. Share your screen or camera in Gemini Live to brainstorm, troubleshoot, and more. Rolling out to Pixel 9 and Samsung Galaxy S25 devices today and available for all Advanced users on Android in the Gemini app:

Introducing Ironwood, the first TPU built for the age of inference, and the timing could not be better : ) - Ironwood perf/watt is 2x relative to Trillium, 6th gen TPU - Ironwood offers 192 GB per chip, 6x that of Trillium - 4.5x faster data access blog.google/products/googl…

Gemini 2.5 Pro just got the final 8th badge in Pokemon Blue, incredible pace of progress by the world's most powerful model!!! Next up: victory road and final 4 : )

Watch Gemini 2.5 Pro implement a landmark Google DeepMind research paper. 🕹️ It codes the reinforcement learning algorithm, visualizes the training live and even debugs errors. ↓

Boom! This model is getting seriously useful.

1. Take a screen recording explaining your app 2. Upload it to YouTube 3. "Build me this” Gemini 2.5’s ability to comprehend video feels straight out of a science fiction novel.

Gemini 2.5 Pro (05-06) is SOTA at most video understanding tasks (by a large margin) 📽️. Lots of work by the Gemini multimodal team to make this happen, excited to see developers push this capability in new ways. More details below!

cooking up something tasty for tomorrow...

Thanks! Of course, when I wrote this blog comment I had also realised that there was a winning move based on this heretical belief: start an AGI company long before most people realise it's coming. Hence, Google DeepMind

I'm a co-author on this new paper on Chain of Thought Monitoring: arxiv.org/abs/2507.11473 It's very related to the System Two Safety idea that I've been talking about for a few years: youtube.com/watch?v=8IUIGV…

What if you could not only watch a generated video, but explore it too? 🌐 Genie 3 is our groundbreaking world model that creates interactive, playable environments from a single text prompt. From photorealistic landscapes to fantasy realms, the possibilities are endless. 🧵

I have to admit, when I saw them generating flakey two second clips of 2D platform games... I did not expect this level of performance just two years later! Wow! Well done Genie Team.

AI efficiency is important. Today, Google is sharing a technical paper detailing our comprehensive methodology for measuring the environmental impact of Gemini inference. We estimate that the median Gemini Apps text prompt uses 0.24 watt-hours of energy (equivalent to watching an

Huge congrats to the Gemini team. Seriously impressive progress to reach this point.