Google releases Gemini 4 Hydrogen for accelerated agentic workflows

  • Gemini 4 Hydrogen
  • Google
  • Agentic workflows

The new model optimizes speed for everyday coding and complex tasks

Features include dynamic model switching and enhanced security protocols

Google introduced Gemini 4 Hydrogen on November 16, 2026, as a high-efficiency artificial intelligence model designed for rapid execution. The model supports developers and consumers by improving performance in coding, testing, and multi-step agentic workflows where agents perform tasks independently. It is available across Google AI Studio, the Gemini API, Gemini for business, the Gemini app, Google Search, and Google Antigravity.

Hydrogen features four distinct thinking levels to balance reasoning depth with cost and latency. It is priced at $1.50 per million input tokens and $7.50 per million output tokens for its 1 million-token context window. To optimize performance, Google also launched Orchestrator in Google Antigravity, which enables dynamic switching between different models during a single session. The model includes upgraded security features to improve resilience against indirect prompt injection and enhance transparency in agentic evaluations.