Google Launches Gemini 4 Hydrogen as New Everyday Workhorse Model, Now Powering Search
The 12 November launch also introduces an Orchestrator that switches between Google models mid-session depending on the task.
Google introduced Gemini 4 Hydrogen on 12 November as the newest member of the Gemini 4 family, positioned as a fast everyday model for coding and agentic workflows. It is available across Google AI Studio, the Gemini API and the Gemini app, and since the same day serves as the model behind Google Search globally - which Google's VP of Search called the biggest model change Search has ever received. Google also launched an Orchestrator in Antigravity that switches between Carbon and Hydrogen mid-session depending on workload. The company reports partner-reported gains of 2x-2.8x speed and 12-23% token savings versus 3.8 Flash on enterprise workloads - vendor figures without independent audit so far.
Two signals stand out for enterprise buyers. First, the family split - frontier model, everyday model, and a router that unifies them - is becoming a standard pricing and quality architecture for AI APIs heading into 2027. Second, with Hydrogen now carrying global Search inference volume, per-token costs for everyday-class models are likely to keep falling, which matters for anyone benchmarking API budgets for drafting, classification and summarization workloads.