Product Launch4 min read

Google releases Gemini 4 Hydrogen, a faster model for agentic work

Google has released Gemini 4 Hydrogen, a 1M-token model priced at $1.50/$7.50 per million tokens that it pitches as the everyday model for coding and agent tasks, alongside an Orchestrator in Antigravity that switches between Hydrogen and Gemini 4 Carbon.

AE
Aidose Editorial
Nov 13, 2026

Google released Gemini 4 Hydrogen on November 12, describing it in a post by Google DeepMind SVP Koray Kavukcuoglu as "our daily driver for agentic workflows." It is built for fast execution on coding and agent tasks and sits alongside Gemini 4 Carbon and Gemini 4 Argon, which Google positions as its frontier models for deep, long-horizon reasoning.

Hydrogen accepts up to 1 million input tokens and can produce up to 1 million output tokens. Google prices it at $1.50 per million input tokens and $7.50 per million output tokens, flat across the full context window. It offers low, medium and high thinking levels, with high as the default, plus a new xhigh tier.

On Google's reported benchmarks, Hydrogen scores 51 on the Artificial Analysis Intelligence Index at its default setting (54 at xhigh), compared with 50 for Gemini 4 Argon. It reaches 57% on Terminal-Bench 4.1 versus 18.2% for Gemini 3.8 Flash, and 71.6% on Agents' Last Exam v1.1. Google says it completes tasks 2.5 times faster than Gemini 3.8 Flash, and that early-access partners saw 2x to 2.8x faster runs with 12% to 23% lower token usage. On safety, the company reports a 12x reduction in misaligned actions and task shortcutting across its tool-use and coding evaluations, and a 96.7% rate on its internal agent honesty evaluation.

The model is available in Google AI Studio and the Gemini API across all tiers, in Android Studio, and in the Gemini app's model picker, with free users receiving it through automatic model selection. It is rolling out in Google Search globally, and in Google Antigravity for paid Google AI Pro and Ultra subscribers.

Google also introduced Orchestrator in Antigravity, which switches between Carbon and Hydrogen within a single session. Google says that on Terminal-Bench 4.0, Orchestrator came close to Carbon's task success while using 24% less quota on average and finishing tasks 19% faster. For developers, the practical change is a cheaper default model with a large output budget for agent loops, and a routing layer that reserves the frontier model for the steps that need it.

AE
Aidose Editorial
Nov 13, 2026 · 4 min read
Back to News