The development
What happened
Google introduced Gemini 4 Hydrogen on 12 November, an AI model aimed at everyday coding and multi-step agent work. The launch spans developer tools, the Gemini app and Search, but access rules differ by product; Google's performance claims are not independent TideReach tests.
Key facts
Listed API pricing is $1.50 per million input tokens and $7.50 per million output tokens, flat across a one-million-token context window.
Developers get low, medium, high (default) and xhigh thinking levels. Google lists API access in AI Studio and Android Studio, plus Gemini for business; its individual-developer Antigravity route is offered to paid Google AI Pro or Ultra subscribers.
Google reports 57% on Terminal-Bench 4.1 versus 18.2% for Gemini 3.8 Flash, with tasks finishing 2.5 times faster at Hydrogen's high setting. These are publisher-reported comparisons, not general speed guarantees.
Antigravity's new Orchestrator can switch models within a session, pairing Hydrogen's execution role with models such as Carbon. This is distinct from Hydrogen operating alone.
Google says Hydrogen is available globally in Search, using query fan-out across question subtopics and adding links, tables and other visual responses. Language coverage remains bounded to supported AI Mode languages.
Free Gemini-app users receive Hydrogen through automatic model selection when appropriate; the announcement does not promise them unrestricted manual selection.
The context
Why it matters
The change puts one model across software-agent and consumer-search workflows, rather than announcing a coding benchmark alone. Teams considering it still need to test their own tasks, tool permissions and failure handling: a benchmark comparison or automatic model switch does not establish reliable completion in their environment.