Google Launches Gemini 4 Hydrogen, Makes It the Default Model in Search's AI Mode

Alphabet's Google released Gemini 4 Hydrogen on Thursday, November 12, rolling it out to the Gemini app and making it the new default model for AI Mode in Google Search. Google's search leaders framed it as a step change: Nick Fox called it a "massive upgrade" and Robby Stein said "the biggest model upgrade to Search yet has landed," citing better intent understanding, broader query fan-out that surfaces a wider range of sites, and more visual answers. Hydrogen also handles the hardest questions in AI Overviews.

In its launch post, Google DeepMind positioned Hydrogen as the fast, everyday tier of the Gemini 4 family, with Gemini 4 Carbon and Argon remaining the frontier deep-reasoning models (Google blog). Google's own figures are aggressive: Hydrogen scores 51 on the Artificial Analysis Intelligence Index versus 50 for Argon at default thinking, reaches 57% on Terminal-Bench 4.1 versus 18.2% for Gemini 3.8 Flash, finishes tasks 2.5x faster, and produces more than 5x fewer fabricated claims (Google blog). These are vendor-reported benchmarks and have not been independently replicated in the coverage so far.

The pricing is the commercially important part for developers. Hydrogen costs $1.50 per million input tokens and $7.50 per million output tokens, flat across its 1 million-token context window, below the $2 and $10 introductory rates for Carbon and Argon (Google blog; FourWeekMBA). It is available through the Gemini API, AI Studio, Android Studio and the Antigravity developer platform.

For GOOGL, the launch matters on two fronts: Search, where AI Mode and AI Overviews are now running on a faster model at Google's own scale, and the developer API market, where a cheaper fast model competes directly for agentic and coding workloads. What to watch is whether third-party benchmarks confirm Google's numbers and whether the AI Mode upgrade shows up in engagement or ad commentary on the next earnings call.