Changelog
Gemini 4 Hydrogen takes over from Gemini 3.8 Flash
Google released Gemini 4 Hydrogen on November 12, and every workspace can now pick it for specialists, workflows, and chats, on our keys with no setup. On Artificial Analysis's Intelligence Index, Hydrogen scores 51 at its default high setting and 41 at medium, against 31 for Gemini 3.8 Flash at high. It writes far fewer tokens per task, so a task often costs less despite the higher token price. Everywhere we chose Gemini 3.8 Flash for you, Hydrogen runs now. Compare it with every other model on the models page.
How it works
- Pricing and window. $1.50 input and $7.50 output per million tokens, and $0.075 per million cached input tokens, at one price across its 1M-token window, with image and PDF input.
- Effort. Hydrogen defaults to high, with low and medium to pick from. In our tests it reasoned before answering at every level.
- Where it replaced 3.8 Flash. Auto's Fast, Balanced, and Max tiers fall back to Hydrogen at high, as do runs on an overloaded Claude or GPT model. Memory updates run on Hydrogen at medium. The hosted chat completion service answers its medium intelligence level with Hydrogen.
- Your own picks stay put. A specialist, workflow, or chat you set to Gemini 3.8 Flash keeps running on it at its current price. Google raises 3.8 Flash to Hydrogen's price on January 1.
- Routes. Google Vertex AI serves it first, with Vercel AI Gateway and OpenRouter behind it. When it's overloaded, runs fall back to Claude Sonnet 5.5, then GPT-6.1 Sol.
Also shipped
- If your Google AI Pro or Ultra plan includes Hydrogen in Antigravity, Token Tunnel offers it on the plan you already pay for, and Auto prefers it on a Gemini subscription.