Thursday, November 12, 2026America’s Trusted News SourceFollow: X · Facebook · Bluesky

USA Pulse News

BREAKING · POLITICS · WORLD · SPORTS
LIVE
BREAKING
News

Google Releases Gemini 4 Hydrogen, Its Fast Model for Coding and AI Agents

Advertisement
Advertisement

Google's Gemini app — Hydrogen is now rolling out  across the Gemini app, Google Search and the Gemini API. (DroidApp)

Google DeepMind launched Gemini 4 Hydrogen on November 12, 2026, the fast, execution-focused member of its Gemini 4 family, built for everyday coding and agentic workflows — jobs where an AI agent carries out multi-step tasks on a user's behalf. "Today, we're introducing Gemini 4 Hydrogen, bringing cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows," wrote Koray Kavukcuoglu, senior vice president of Google DeepMind and Google's chief AI architect.

Hydrogen is adapted from Gemini 4 Carbon, the frontier model built for deep, long-horizon reasoning. Carbon and its sibling Argon keep the hardest reasoning work, while Hydrogen is the workhorse: writing code, calling tools, reading documents and running agents at high volume. It ships with four thinking levels — low, medium, high as the default, and a new xhigh tier — so developers can dial reasoning depth up or down for each request, paying and waiting only for what each task needs.

Hydrogen is built for everyday coding and agentic workflows, where an AI agent carries out multi-step tasks. (Hashnode)


On Google's own figures, Hydrogen scores 51 on the Artificial Analysis Intelligence Index at its default high setting, ahead of Argon at 50, and reaches 54 at xhigh. Google says tasks finish 2.5 times faster than with Gemini 3.8 Flash, Terminal-Bench 4.1 — a test of real coding tasks in a terminal — jumps from 18.2 percent to 57 percent, and fabricated claims fall by more than five times. Those are vendor-reported figures, not independently reproduced: one independent check, by orcarouter.ai on November 13, confirmed the intelligence score but found the 2.5-times speed claim did not hold up outside Google's own test setup.

Pricing is flat across the model's one-million-token context window: $1.50 per million input tokens and $7.50 per million output tokens, with no surcharge for long inputs. Google's Orchestrator, which routes work between models, reports near-Carbon results with 24 percent less quota and 19 percent faster completion when Hydrogen handles the execution steps.

Google says Hydrogen finishes tasks 2.5 times faster than Gemini 3.8 Flash — a vendor-reported figure one independent check disputed. (Beehiiv)


Hydrogen is rolling out globally from November 12 across all tiers of Google AI Studio, the Gemini API, Gemini for business, the Gemini app and Google Search for all users, plus Antigravity for paid Google AI Pro and Ultra subscribers. Free Gemini app users get it through automatic model selection rather than choosing it directly, and Gemini in Sheets starts with eligible business subscriptions and AI Pro and Ultra subscribers. The model is also listed on Vercel AI Gateway as google/gemini-4-hydrogen.

Google's headquarters campus in Mountain View, California. (Substack)


Google's own model card notes the limits: a March 2026 knowledge cutoff, occasional hallucinations and slowdowns, and higher token use at the deeper thinking levels. On Google Cloud, Hydrogen is classified as short-term generally available, with at least 45 days' retirement notice and no retirement date announced. The launch caps a rapid Gemini 4 rollout — Carbon reached paid developers and enterprises on November 10 — and sets up a two-tier lineup: Carbon and Argon think, Hydrogen executes.

Sources

— Google DeepMind / Google Keyword Blog ("Introducing Gemini 4 Hydrogen")
— AI Primer
— AI Mastery
— AI Agent Update
— PakTechStan
— APIYI

Advertisement