Signal

Google says Gemini 4 Hydrogen is available across AI Studio, the Gemini API, Search, paid AI Pro and Ultra tiers in Antigravity, and more

First reported by 9to5Google ·

The signal ●●●● Compiled by AI from 9to5Google, Techmeme, Google and RuntimeWire
Why you might care

Google's Gemini 4 Hydrogen is now faster and reduces errors, meaning your AI-generated content and analysis will be more accurate and quicker to produce.

What happened

Google has announced the release of Gemini 4 Hydrogen, a new AI model designed to offer significant improvements in both speed and intelligence. Compared to its predecessor, Gemini 3.8 Flash, Hydrogen performs tasks up to 2.5 times faster and achieves higher benchmark scores in various AI evaluations. It boasts a native context window of 1 million input and 1 million output tokens, and it has demonstrated a reduction in fabricated claims by over five times. Hydrogen also shows enhanced adaptability when encountering tool failures or empty search results. The model is being integrated across Google's products, including the Gemini app, Google Search's AI Mode, and Google Antigravity, where it will serve different tiers of users and assist in complex tasks. Developers will see a per-token cost of $1.50 for input and $7.50 for output.

What it means

Gemini 4 Hydrogen represents a substantial upgrade over previous models, particularly in its ability to handle complex, multi-step tasks and its reduced tendency for errors or overreach. Its integration into products like Google Search and the Gemini app aims to make AI interactions more intuitive and reliable for end-users, while the pricing structure for developers suggests a continued focus on monetizing advanced AI capabilities. The introduction of Orchestrator in Antigravity, which dynamically switches between models like Carbon and Hydrogen, signals a move towards more sophisticated, cost-efficient AI workflows that leverage specialized models for different stages of a task.

The broader implications of Gemini 4 Hydrogen's release point to an escalating race in AI performance and reliability. By offering a model that is faster, generates fewer errors, and has a significantly larger context window, Google is pushing the boundaries for what developers can build and what consumers can expect. This upgrade will likely influence competitive offerings, forcing other AI providers to accelerate their own model development and deployment cycles to maintain market relevance, especially in areas like complex reasoning and agentic task execution.

AI-written summary. May contain errors.