Google releases Gemini 4 Hydrogen, a fast model at $1.50 per million input tokens

Google released Gemini 4 Hydrogen on 12 November 2026, the fast model of the Gemini 4 family and the successor to Gemini 3.8 Flash. It costs $1.50 per million input tokens and $7.50 per million output tokens (a token is a short piece of a word), and Google says the price stays the same across the full 1-million-token context. It is in the Gemini API, the Gemini app and Google Search.

Artificial Analysis lists it at 51 on its Intelligence Index at the default “high” setting and 54 at “xhigh”, a new fourth thinking level. Gemini 4 Argon scores 50 and Gemini 4 Carbon 60 on the same index. Gemini 3.8 Flash scored 31. Artificial Analysis puts the cost of running its test suite at $1.87 per task for Hydrogen (high), against $9.95 for Argon and $7.76 for Carbon.

Google reports its own results too. It says Hydrogen finishes tasks 2.5 times faster than Gemini 3.8 Flash at the default setting and scores 57% on Terminal-Bench 4.1, against 18.2% for 3.8 Flash.

What it means

Google’s fast tier now reaches the index score of Gemini 4 Argon, announced as its flagship six weeks earlier. Argon costs $2 and $10 per million tokens until 30 November and $4 and $20 from 1 December, Google’s pricing page shows. Carbon, Google’s current best, is still nine points ahead.

Caveats

• The speed and Terminal-Bench 4.1 figures are Google’s own. The post links a model card but not the runs behind them.
• One index point separates Hydrogen from Argon, and Artificial Analysis shows whole points only. That gap is too small to say either model is better.
• Cost per task is Artificial Analysis’s measurement on its own tests. Your own jobs can use more or fewer tokens.
• Epoch AI’s Capabilities Index has no score for Hydrogen yet, so it has no rank on the site’s main measure.
• Google also has a free tier and a batch price of $0.75 and $3.75. OpenRouter lists other routes at other prices.