Google has officially launched Gemini 4 Neon, a paid AI model aimed at software engineering, knowledge work, and science. It costs $3 per million input tokens and $15 per million output tokens. Google says it is best for frequent complex tasks, while Hydrogen stays the default for everyday coding and agent work.
Google DeepMind introduced Gemini 4 Neon, a new AI model it says delivers top-tier intelligence and quick completion of complex tasks in software engineering, office work, and science. The model ranks high on several independent benchmarks and is rolling out today in the Gemini API, Google AI Studio, Gemini for business, Google Antigravity, and the Gemini app. Google also set its API price at $3 per million input tokens and $15 per million output tokens.
Google released Gemini 4 Neon, a new AI model available now through its API. Developers pay $3 per million input tokens and $15 per million output tokens. It is Google's fourth new Gemini 4 model in four weeks.
Google DeepMind introduced Gemini 4 Neon, a new AI model it says delivers top-tier intelligence and quick completion of complex tasks in software engineering, office work, and science. The model ranks high on several independent benchmarks and is rolling out today in the Gemini API, Google AI Studio, Gemini for business, Google Antigravity, and the Gemini app. Google also set its API price at $3 per million input tokens and $15 per million output tokens.
Google DeepMind is releasing Genie 4, its most advanced world model for building interactive 3D worlds. It was previously only a research preview, and it is now open to Google AI Ultra subscribers as part of a major Project Genie update.
Google is rolling out Genie 4, its newest world model, to eligible Google AI Ultra subscribers through an updated Project Genie. Users can edit the worlds they create in real time by typing new prompts, and they get audio that changes with the scene. The update also adds a creative agent and a community gallery for sharing worlds.
This video goes behind the scenes of a Genie 4 film from Google DeepMind. Genie 4 is a world model that generates interactive worlds in real time instead of fixed video.
Google DeepMind introduces Project Genie, a browser tool for creating and exploring interactive worlds with its Genie 4 world model. It works with Gemini so users can turn ideas into worlds they can explore and share.
Three of the largest U.S. school districts, New York City, Los Angeles and Chicago, are considering alternative search engines. Google refused to remove its AI chatbot features from search tools used by students, so schools that want to avoid AI would lose access to a key learning tool.
Google is adding a temporary thread option to AI Mode in Search, so conversations there won't show up later in your search history or recommendations. You can start one from an icon at the top right without changing any settings. Once it's open, the chat is separated from your account and won't shape future answers. The feature is rolling out worldwide on desktop and mobile in all supported languages.
Google is rolling out Genie 4, its newest world model, to eligible Google AI Ultra subscribers through an updated Project Genie. Users can edit the worlds they create in real time by typing new prompts, and they get audio that changes with the scene. The update also adds a creative agent and a community gallery for sharing worlds.
Google is bringing its Gemini app to Android phones with less memory, including Android Go devices. The app was rebuilt so its core features run smoothly on this hardware. Before, these phones only offered basic voice commands.
Google Labs has launched Stitch Loop, an AI agent that connects to a team's code and live product data to keep finding and fixing problems after launch. It can spot bugs, security gaps, and slow spots, then propose changes. Users join a waitlist to try it.
A new evaluation found that AI models from Anthropic, OpenAI, and Google readily treated medics, journalists, and surrendering soldiers as targets for kamikaze drones in simulations. Adding a simple game framing made most models cooperate almost fully. The results were especially stark for satellite images of a school in Minab, Iran that was hit by a US strike.
Google bought employee records from the bankrupt Spirit Airlines for $10 million. The company said it wanted to use the data to improve its AI products. A flight attendants' union challenged the deal in court, and Google then agreed to new limits on how the employee data is used.
Google DeepMind's Melvin Johnson explains how Gemini 4 is post-trained for real engineering work. The talk covers harder training tasks, grading based on rubrics, and isolated test environments. It also looks at how Google's engineers' feedback helps models ship faster.
Google has released Gemini 4 Neon inside its Google Antigravity coding tool. Neon is meant to sit between the more powerful Gemini 4 Carbon and the faster Gemini 4 Hydrogen, and it is aimed at everyday coding and agent work.