Google launches Gemini 4 Hydrogen as its fast everyday model, replacing the Flash line
Hydrogen is priced at $1.50 per million input tokens and is rolling out in the Gemini app, Search and Antigravity, where a new Orchestrator pairs it with Carbon.

- What happened
- Google released Gemini 4 Hydrogen on 12 November, a faster, cheaper Gemini 4 model aimed at everyday coding, agentic tasks and consumer use.
- What’s new
- Hydrogen succeeds Gemini 3.8 Flash under a new name and adds a fourth, higher thinking level plus an Orchestrator that mixes it with Carbon.
- Why it matters
- Hydrogen brings Gemini 4 to free Gemini app users and all Search users, and gives developers a cheaper execution model for multi-model workflows.
Google introduced Gemini 4 Hydrogen on 12 November, positioning it as the quick, low-cost member of the Gemini 4 family alongside the larger Gemini 4 Carbon and Gemini 4 Argon. Google describes Carbon and Argon as its frontier models for long-horizon reasoning, while Hydrogen is meant for rapid execution of everyday coding and agentic work. 9to5Mac notes that Google dropped the Flash branding for this model and that it arrived days after Carbon became available.
For developers, Google prices Hydrogen at $1.50 per million input tokens and $7.50 per million output tokens, flat across its 1 million-token context window. The model accepts 1 million input tokens and can produce up to 1 million output tokens.
Claimed gains over Gemini 3.8 Flash
Google pitches Hydrogen as an upgrade for teams on Gemini 3.8 Flash, which 9to5Mac says launched in September. According to Google, Hydrogen finishes tasks 2.5 times faster at its default high setting, scores 57% on Terminal-Bench 4.1 against 18.2% for 3.8 Flash, and cuts fabricated claims by more than five times. Google also says Hydrogen scores 51 on the Artificial Analysis Intelligence Index at its default level, edging Argon's 50, and reaches 54 at its highest setting.
Some of these figures come from limited testing. Google says partners saw Hydrogen run 2x to 2.8x faster with 12% to 23% lower token use than 3.8 Flash, but describes this as early testing across graph, SQL and data-heavy enterprise workloads. The benchmark claims, including class-leading results on HLE-Diamond and the Vals tax and legal benchmarks, are Google's own reporting. On safety, Google says Hydrogen cuts misaligned actions and task shortcutting by 12x relative to 3.8 Flash in tool-use and coding evaluations.
Orchestrator and thinking levels in Antigravity
Alongside the model, Google added Orchestrator to its Antigravity coding tool. Orchestrator switches between models such as Carbon and Hydrogen within a single session; in parallel-agent setups, Carbon coordinates while Hydrogen runs worker and verifier agents. Google says that on Terminal-Bench 4.0 Orchestrator came close to Carbon's success rate while using 24% less quota on average and finishing 19% faster.
Google's announcement describes four thinking levels for Hydrogen: low, medium, the default high, and a new xhigh tier. 9to5Mac reports that the Antigravity interface labels the top tier Extra. Antigravity access is limited to paid Google AI Pro and Ultra subscribers.
Who gets Hydrogen and how
Google says Hydrogen is rolling out globally in the Gemini API, AI Studio, Gemini for business, the Gemini app, Google Search and, for eligible subscribers, Gemini in Google Sheets. In the Gemini app's model picker, 9to5Mac reports, Hydrogen is labelled for everyday help, alongside 3.5 Flash-Lite, 3.1 Pro, 4 Argon and 4 Carbon. According to 9to5Mac, free app users reach Hydrogen through automatic model selection when it suits the prompt. In Search, Google says Hydrogen should better interpret multi-part or typo-laden queries, give more visual answers and surface a wider range of links.