Gemini 4 Hydrogen explained: price, speed and access.

Gemini 4 Hydrogen is Google’s new fast model in the Gemini 4 family, launched on 12 November 2026. Google says it brings “cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows”, and that with Hydrogen it is “bringing Gemini 4 to everyone”, including all tiers of Google AI Studio and the Gemini API, the Gemini app and Google Search. Here is what Gemini 4 Hydrogen is, what is new, what it costs and how to try it.

What is Gemini 4 Hydrogen?

Google positions Hydrogen next to its frontier models, Gemini 4 Carbon and Gemini 4 Argon. Google says Carbon and Argon “excel at deep, long-horizon reasoning on your hardest problems”, while Hydrogen “is built for rapid execution across diverse agentic workflows”. It can be used on its own for everyday work, or paired with Carbon and Argon as what Google calls “the high-efficiency execution engine in multi-model workflows”.

For developers, Google describes Hydrogen as the upgrade path from Gemini 3.8 Flash. The model ID in the Gemini API is `gemini-4-hydrogen`.

What’s new in Gemini 4 Hydrogen

According to Google’s announcement and Gemini API docs:

• Smarter than Argon on one index. Google says Hydrogen scores 51 on the Artificial Analysis Intelligence Index at its default high thinking level, versus 50 for Argon, and reaches 54 at its highest `xhigh` setting.
• Faster than 3.8 Flash. Google says it finishes tasks 2.5 times faster than Gemini 3.8 Flash at its default setting, more than triples the Terminal-Bench 4.1 score (57% vs. 18.2%) and cuts fabricated claims by more than 5x.
• Four thinking levels. `low`, `medium`, `high` (the default) and a new `xhigh` tier. Google’s docs say `minimal` is rejected with an error.
• 1M-token context with long outputs. The docs say input, thinking and output share one 1,048,576-token window, with a 65,536-token default per request and longer outputs (up to 1M tokens) via background execution or SDK continuation.
• Multimodal input. Google says Hydrogen can work over PDFs, spreadsheets, charts, audio and recorded video to produce cited reports and structured deliverables.

Google also lists benchmark results for agentic and knowledge work: 58.8% on HLE-Diamond, 52.3% on Vals Tax Agent Bench (ranked #1 overall), 55.8% on the Vals Legal Research Bench (tied for #1), 53.4% on the new Multimodal Knowledge Agent (MKA) benchmark and 71.6% on Agents’ Last Exam v1.1. Google says it ranks #5 on arena.ai’s Agent Arena. These are Google’s own reported figures.

Gemini 4 Hydrogen price

Google says Gemini 4 Hydrogen is priced flat across its full 1M-token context window at $1.50 per million input tokens and $7.50 per million output tokens on the paid tier. It is also available on the free tier of Google AI Studio and the Gemini API. The docs add that implicit caching runs by default on prompts of at least 4,096 tokens and discounts cached input tokens by 95% on cache hits, while explicit context caching is not supported.

Orchestrator: Hydrogen and Carbon together

Alongside Hydrogen, Google introduced Orchestrator in Google Antigravity, which switches between models like Carbon and Hydrogen within the same session. Google says that on Terminal-Bench 4.0, Orchestrator reached near-Carbon task success while using 24% less quota on average and finishing 19% faster. For parallel agents, Google suggests Carbon as the coordinator with Hydrogen powering worker and verifier agents.

How to try Gemini 4 Hydrogen

Google says Hydrogen is rolling out globally from 12 November 2026:

• Developers: the Gemini API in Google AI Studio (free and paid tiers) and Android Studio, plus Google Antigravity for paid Google AI Pro and Ultra subscribers.
• Businesses: Gemini for business via the API or model drop-down, Google Antigravity, and Gemini in Google Sheets for eligible business subscriptions.
• Everyone: the Gemini app model drop-down and Google Search for all users. Google says free Gemini app users get Hydrogen through auto model selection when it fits the prompt.

If you are moving code from Gemini 3.8 Flash, Google’s migration notes boil down to a few steps:

1. Change the model ID to `gemini-4-hydrogen`.
2. Raise or remove low `max_output_tokens` caps, since input, thinking and output now share one 1M-token window.
3. Set `thinking_level` explicitly: Hydrogen defaults to `high`, not `medium`, so set `medium` if you want 3.8 Flash-like latency.
4. Remove `temperature`, `top_p`, `top_k` and explicit caching fields, which Google says are ignored or unsupported.
5. Declare tools natively with function calling or use structured outputs instead of asking for raw-text JSON tool calls.

What people are building with Gemini 4 Hydrogen

The first Hydrogen builds showed up on X within hours of launch. Three are on The Model Atlas so far:

• Neon city flythrough built with Gemini 4 Hydrogen: a Three.js city you can fly through, with procedural towers, rain and flying traffic, made with Hydrogen High in Antigravity.
• Island lagoon with a sailboat: an interactive 3D lagoon built with Hydrogen and Antigravity’s Swarm feature, with water shaders, changing weather and a sailboat you can steer.
• Light a star, a fusion game: a browser game Hydrogen built from one prompt in 78 seconds that walks you through the proton-proton chain inside a young star.

See the Gemini 4 Hydrogen model page for every build we add, and our Gemini 4 Carbon guide for the frontier model it pairs with.

Who is Gemini 4 Hydrogen for?

Going by Google’s description, Hydrogen is for developers and teams who want a fast, cheaper model for everyday coding, debugging, data work and agent tasks, and for anyone using Gemini or Search day to day. Teams that need deep reasoning on the hardest problems can keep Carbon or Argon for planning and hand execution to Hydrogen.