Particle.news

Google Rolls Out Gemini 4 Hydrogen as Fast, Low-Cost Model for Agentic Workflows

The model promises faster, lower-cost execution for multi-step coding and task workflows.

Overview

  • Google began a global rollout of Gemini 4 Hydrogen on Friday, November 13, 2026, making it available in Google Search, the Gemini app, the Gemini API and AI Studio.
  • Hydrogen runs across a 1 million‑token context and is priced at $1.50 per million input tokens and $7.50 per million output tokens, creating new cost tradeoffs for long, multi-step sessions.
  • Google says Hydrogen is roughly 2.5 times faster than Gemini 3.8 Flash and positions it as a workhorse for everyday tasks such as writing code, running tests and executing agentic workflows.
  • Company-published benchmarks show Hydrogen outscoring some heavier Gemini models and rivals on selected tests, but those results and Google’s performance claims lack independent verification.
  • Google highlights Antigravity orchestration that shifts work between Carbon and Hydrogen to cut quota use and speed runs, and it warns that Hydrogen still faces safety limits such as hallucinations and jailbreak risks.