Google on Nov. 12, 2026, introduced Gemini 4 Hydrogen, a new artificial intelligence model designed for coding and automated workflows.
The system is rolling out globally to developers, enterprises, and consumers. It is available across Google platforms, including Google Search, Google AI Studio, and the Gemini application.
The company set the price at $1.50 for every million tokens users input and $7.50 for each million the system produces, across its 1-million token capacity. The model features four thinking levels—low, medium, high, and a new xhigh tier—which allow users to scale quality, latency, and cost based on task requirements, according to the company. The high tier is the default setting.
The company said Hydrogen finishes tasks 2.5 times faster than its previous Gemini 3.8 Flash model at its high default setting. Google also stated the new model cuts fabricated claims by more than five times compared to the earlier version.
Alongside the model, Google introduced Orchestrator in Google Antigravity, a tool that automatically switches between different models during a single session. On a benchmark test, Orchestrator used 24% less quota and finished tasks 19% faster on average, the company said. Google also stated that Hydrogen reduces overreach in automated tasks by up to 81% and cuts misaligned actions across coding evaluations by 12 times.