Vibitely Daily

Friday, 13 November 2026, story 1 of 7

Google launches Gemini 4 Hydrogen to replace Flash, edging out Argon

Google says it scores one point above Argon on the Artificial Analysis index, at $1.50 per million input tokens and $7.50 per million output.

Illustration: A small, light flask-shaped vessel sits on a balance scale and tips it slightly past a larger, heavier vessel on the other pan, while a third tall vessel stands apart behind them.

What happened

Google released Gemini 4 Hydrogen on Nov 12. It is available now in the Gemini API, Gemini for business, the Gemini app and Search, and for paid Google AI Pro and Ultra subscribers in Antigravity. The Decoder reports it is meant to replace Gemini 3.8 Flash. It costs $1.50 per million input tokens and $7.50 per million output tokens, flat across a 1 million-token context window.

Google says Hydrogen scores 51 on the Artificial Analysis Intelligence Index at its default high setting, against 50 for Argon, and 54 at the new xhigh level. It also reports 57% on Terminal-Bench 4.1 versus 18.2% for 3.8 Flash.

What it means

The lead over Argon is one index point, and the figure comes from Google. As The Decoder reports it, Artificial Analysis's own evaluation puts Hydrogen at 51 on high, behind GPT-6.1 Sol and Claude Opus 5.5 at 52 each. The Decoder's heading also says an independent test found it lags on terminal work, though the excerpt gives no figures. Agentic coding is therefore the place to test hardest.

Teams that settled on Argon for daily coding should rerun their own tasks on Hydrogen. The Decoder says Google recommended Argon for that work only two days earlier, so the guidance is likely to shift. Carbon stays Google's pick for hard problems. Cost per task will depend on which thinking level you choose and how many output tokens it burns, so measure both. The source does not say whether Orchestrator, Google's model-switching feature, is available outside Antigravity.

What to do: This week, run your own coding and agent tests on Hydrogen at high and xhigh against your current Argon or Flash setup, tracking both success rate and tokens spent.

Sources

Filed under Models. Published .

"What it means" is analysis, not reporting. Found an error? Write to daily@vibitely.com.