Skip to content
Trending
BreakingAI1 min read

Gemini 4 Hydrogen is live: free in Search, $1.50/million input tokens for devs

Free in Search and the Gemini app. $1.50/million input tokens for developers. Live now.

By Inference

The AI Desk · (5 hours ago)

Reported fromoutlets, credited2
Google Gemini logo with 'Hydrogen' badge
Google Gemini logo with 'Hydrogen' badgePhoto via The Next Web

Gemini 4 Hydrogen is live — and it’s free in Search and the Gemini app. Google launched the model globally on Thursday, pricing it at $1.50 per million input tokens for developers, with output tokens at $7.50 per million. That’s less than half the cost of Gemini 4 Argon, its flagship model.

In the US and India, no new subscription or tier is required: anyone using Google Search or the Gemini app gets Hydrogen by default, no sign-up, no waitlist, no credit card. No price is listed for India specifically, but the $1.50/million input token rate applies to developers worldwide, per Google’s announcement.

The model handles up to 1 million input tokens at once, same as Argon, and Google says it hallucinates five times less often than Gemini 3.8 Flash. It’s built for coding, agent workflows, and messy, multi-part search queries, the kind that trip up older models by mixing intent, context and ambiguity.

In Search, it pulls from a wider set of sources than before, and links more broadly across the web. Independent tester Artificial Analysis ranks Hydrogen one point above Argon at default settings, though it still sits one point below Anthropic’s Claude Sonnet 5.5 at its top setting. Vals AI calls it the cheapest and fastest model in its top six.

Google also rolled out Orchestrator inside Antigravity, its coding assistant: it dynamically swaps between Hydrogen and Gemini 4 Carbon, hitting 96% of Carbon’s output quality while using 24% less quota. That’s not a replacement for Carbon, it’s a throttle. Hydrogen does the grunt work. Carbon plans.

And yes, it’s 2.5× faster than 3.8 Flash. But OpenAI just cut GPT-6 pricing, and Claude Sonnet 5.5 remains slightly ahead on benchmarks. So this isn’t about winning every test. It’s about making high-speed, low-cost reasoning available to everyone, not just paying devs.

Free in Search. Free in the app. Live now.

Reported from

How this story was made

Written by the Hitechreports desk from the reporting credited above, with facts attributed to their original publishers. We do not test devices ourselves; anything about performance, battery life or cameras comes from the outlets that did. Prices are as reported at the time of writing. Editorial policy · Report an error

What did you make of it?

Filed by The AI Desk

Models, assistants and the companies and chips behind them, reported from what was released and what was claimed, with the difference kept clear.

More from this desk →

Be the first to comment

Join the argument. No password, just your email or a passkey.

    Read next

    More AI →