Welcome to the Gemini 3 era ⚡
Tulsee Doshi, Senior Director of Product Management for Gemini, breaks down the latest on our latest models Gemini 3 and Nano Banana Pro ⬇️
Hey everyone, I'm Chelsea and I lead product for our Gemini models. Here are three things you need to know about our Gemini 3 release. One thing I'm really excited for everyone to experience is Gemini's state-of-the-art intelligence and reasoning. It really operates with a level of depth and nuance. It is leading on a number of benchmarks that measure academic reasoning like humanities last exam or GP QA. So it's really able to have that kind of PhD level knowledge. It also just has an improved persona. It drops the cliches. And the over flattery and is noticeably better to interact with. So my personal favorite part of this Gemini release is it's multimodal understanding. It is incredibly capable of understanding video, image and actually even more than that, combining modalities, whether that is audio or image or code or text and bringing these things together, you can create rich, truly beautiful visuals that can actually help you understand any concept. And Gemini is the best vibe and agentic coding model we've ever built. It tops the web dev arena leaderboard with the score of 1487. And what that really means is it can just build these incredibly rich and visual web apps with even simple prompts. I actually used it earlier this morning to build a web app to help me with dance choreography and formations, which was awesome. For the developers out there though who want a proper IDE, we actually also just released Google Anti Gravity, which is. Our take on the agentic coding experience. We also introduced Gemini 3 Pro image or Nano Banana Pro, which is our state-of-the-art, honestly incredible image generation and editing model. This model has a next level of precision, which means it's amazing at text rendering and control, which makes it super valuable for pros and production ready visuals. And because it's built on Gemini 3 Pro, it's actually really able to use Geminis real world knowledge. And so that makes it not only really good at rendering text, but also at localizing to different languages and building on physics to be able to actually build aesthetically beautiful but also realistic imagery. Gemini 3 is already available across a number of our Google products. You can try it today at DeepMind dot Google slash Gemini.
Building I.N.G: A real-time video platform where your curiosity becomes income. Let’s connect.
(Your curiosity deserves ROI.)| Real-time platform | Shortform | AI video tech
Educator & Learning Facilitator | Transitioning to Corporate Training, Learning & Development | M.A. in Curriculum & Instruction
16hI can’t wait to try out Gemini 3 along with the new image generation upgrades. Google is doing exciting things with AI!