Google has launched Gemini 4 Hydrogen, a lightweight general-purpose AI model designed to handle everyday tasks quickly and cheaply across developer APIs and consumer applications. Despite its smaller positioning compared to Google’s frontier Carbon and Argon models, Hydrogen scored 51 at its default setting and up to 54 at its highest tier on the Artificial Analysis Intelligence Index, outperforming the larger Argon model released less than a month prior.

Hydrogen features a one-million-token context window for both input and output, priced at $1.50 per million input tokens and $7.50 per million output tokens. According to Google, the model operates 2.5 times faster than Gemini 3.8 Flash, achieves a 57 percent score on Terminal-Bench 4.1, and reduces hallucinations by 5x while offering four distinct thinking levels.

The model is rolling out globally across Google Search, the Gemini app, Gemini API, Android Studio, and developer environments like Antigravity. In Antigravity, Google has introduced a dynamic orchestrator that automatically switches between Carbon and Hydrogen within a single session to optimize coding efficiency.

Why it matters

  • Developers gain access to frontier-class reasoning performance at mid-tier pricing ($1.50/$7.50 per M tokens).

  • Fast model release cycles are rendering recent frontier releases obsolete within weeks, accelerating deprecation schedules.

  • Dynamic orchestrators switching between cheap and heavy models within single developer sessions represent the new standard for AI tooling.

Source: the-decoder.com