Google introduced Gemini 4 Hydrogen on Thursday, a model aimed at everyday work: writing code, running tests and agentic workflows. According to the company, the larger Gemini 4 Carbon and Argon models remain for problems that need long reasoning, while Hydrogen is meant to execute faster and cheaper. Unlike Carbon and Argon, which rolled out gradually starting with security researchers, Hydrogen is available immediately and replaces Gemini 3.8 Flash across Google's products.
According to Google's figures, at its default (high) thinking level the model finishes tasks 2.5 times faster than 3.8 Flash, scores 57% on Terminal-Bench 4.1 versus 18.2%, and cuts fabricated claims by more than 5x. On the Artificial Analysis Intelligence Index it scored 51 at the default level and 54 at the new xhigh level, versus 50 for Argon. It has four thinking levels: low, medium, high and xhigh.
API pricing is flat across the full one-million-token context window: $1.50 per million input tokens and $7.50 per million output tokens, the same as 3.8 Flash. According to Ars Technica, Carbon and Argon pricing will rise next month from $2/$10 to $4/$20. Google also introduced Orchestrator in its Antigravity development environment, which automatically switches between Carbon and Hydrogen within the same task.
Hydrogen is also coming to Google Search, in AI Overviews and AI Mode, for all users. In the Gemini app it is available to AI Pro and Ultra subscribers, and free users will get it through automatic model selection when it fits the prompt. Google has not yet published a full comparison with rival models beyond the figures it chose to present.
