Google’s Gemini 4 Hydrogen runs 2.5x faster than 3.8 Flash and edges past Argon

Gemini 4 Hydrogen was announced on November 12, 2026, and is immediately rolling out across all Gemini tiers. The model is live in the Gemini app, the Gemini API, Google AI Studio, Vertex AI and Google Search. Its most notable achievement is completing tasks 2.5 times faster than Gemini 3.8 Flash, while also outperforming Google's own frontier model, Gemini 4 Argon, on the Artificial Analysis Intelligence Index.

Performance benchmarks

Hydrogen scored 51 on the Artificial Analysis Intelligence Index at default settings, one point above Argon's score of 50. When pushed to maximum capacity, Hydrogen's score climbs to 54. The most dramatic improvement came in coding performance, where Hydrogen posted a score of 57% on Terminal-Bench 4.1, up from just 18.2% in its previous version.

Terminal-Bench measures how well a model handles tasks in a command-line environment, the text-based workspace developers use to run code and manage systems. This significant jump suggests Hydrogen could be particularly valuable for teams building coding assistants and automated agents.

Positioning and pricing

Google is positioning Hydrogen as the fast, practical member of the Gemini 4 family, designed for rapid execution, algorithm coding and general-purpose work. The heavier computational tasks remain with Gemini 4 Carbon and Argon, the more capable frontier models. Hydrogen is built for multi-modal orchestration alongside these models, allowing them to split workloads rather than compete.

The model carries a context window of 1 million tokens and is priced at a flat rate of $1.50 per million input tokens and $7.50 per million output tokens. This pricing structure is aimed at developers and high-volume users, offering predictable costs despite the five-to-one gap between output and input pricing.

Testing and implications

Hydrogen's development included community testing from late October 2026 under the name gemini-3.8-flash, followed by internal staff testing reported on November 11, 2026, one day before the public announcement. This quiet testing run allowed Google to refine the model before its official release.

For developers, the combination of speed, coding improvements and predictable pricing could significantly change how responsive AI-powered applications feel. The model's performance raises questions about Google's own lineup, as the fast tier now scores higher than the frontier model at default settings. Google's solution is orchestration, positioning Hydrogen to work alongside other models rather than replace them.