Google Gemini 4 Hydrogen release: multi-part search and 1M context
Google released Gemini 4 Hydrogen on 12 November 2026. It is a fast general-purpose model that replaces Gemini 3.8 Flash across Search, the Gemini app, Google AI Studio, the Gemini API, Gemini for business, and Google Antigravity for AI Pro and Ultra subscribers. Google says the model handles questions with multiple parts, hard-to-understand typos, and highly specific requests.
For teams that ship software or run small delivery operations, the relevant points are speed, price, and the new xhigh thinking level. The API price stays at $1.50 per million input tokens and $7.50 per million output tokens, the same as Gemini 3.8 Flash.
In one line: Google released Gemini 4 Hydrogen, a faster general-purpose model that replaces Gemini 3.8 Flash, handles complex multi-part search queries better, and adds four thinking levels including xhigh.
Availability, pricing, and context window
Gemini 4 Hydrogen is available now in Google Search, the Gemini app, all tiers of Google AI Studio, the Gemini API, Gemini for business, and Google Antigravity for AI Pro and Ultra subscribers. Ars Technica notes that in the Gemini app and AI Studio, AI Plus and free users remain on Gemini 3.5 Flash-Lite, so not everyone gets the model immediately.
The model has a 1 million token context window for both input and output. Google lists flat pricing of $1.50 per million input tokens and $7.50 per million output tokens. Cache hits cost $0.075 per million tokens according to The Decoder. Hydrogen has four thinking levels: low, medium, high default, and a new xhigh tier.
Benchmarks and independent tests
Google says Hydrogen surpasses Gemini 4 Argon on the Artificial Analysis Intelligence Index at the default high level, 51 versus 50, and reaches 54 at xhigh. The Decoder reports that Artificial Analysis puts xhigh Hydrogen roughly on par with Claude Sonnet 5.5. On the AA-Omniscience benchmark, Hydrogen made things up only 8 percent of the time. It also beat other Gemini models, including Carbon, on presentation quality in AA-Briefcase.
Google reports Hydrogen more than triples Terminal-Bench 4.1 scores compared with Gemini 3.8 Flash, 57 percent versus 18.2 percent. It ranks fifth on Agent Arena, first on Vals Tax Agent Bench at 52.3 percent, ties for first on Vals Legal Research Bench at 55.8 percent, and scores 53.4 percent on the Multimodal Knowledge Agent benchmark and 71.6 percent on Agents' Last Exam v1.1.
Independent tests also show limits. Artificial Analysis measured Hydrogen at 54 percent on Terminal-Bench 4.0 and 50 percent on Terminal-Bench-Science 0.1. Claude Sonnet 5.5 max reached 64 and 53 percent, and GPT-6.1 Sol max reached 56 and 58 percent. At xhigh, Hydrogen costs about $3.16 per task, roughly a third of Claude Sonnet 5.5 max but about 2.5 times as much as GPT-6.1 Sol max according to Artificial Analysis.
Search and Gemini app changes
Google says Hydrogen is the biggest model upgrade to Search yet. Robby Stein, VP of product at Google Search, told The Verge that Hydrogen is now the default model in AI Mode and is used for the hardest questions in AI Overviews. The model can break down a query and address each part. It also runs simultaneous queries across relevant subtopics, which Google says produces a wider range of links. Responses include more visuals like images, comparison tables, graphs, and maps. Google says language understanding improved across all supported AI Mode languages.
In the Gemini app, Hydrogen follows tone, format, and instructions more closely over multi-turn conversations. Google says email drafts, study guides, rewrites, and documents are ready to use on the first try more often. The model pulls context from uploads, Gmail, Google Drive, and connected apps like Airtable and Dropbox. If you rely on Search AI Mode or app assistants, the research and search tools directory lists other options to compare.
Developer workflows and agentic coding
Google positions Hydrogen as a standalone upgrade for teams moving from Gemini 3.8 Flash. Google says it finishes tasks 2.5 times faster at the high default and cuts fabricated claims by more than 5 times. For repository work, Hydrogen handles issue resolution, debugging, rapid iteration, and codebase Q&A with focused edits. It batches file read and search calls in parallel, which Google says helped partners run 2 to 2.8 times faster and cut token usage by 12 to 23 percent compared with 3.8 Flash.
For multi-model workloads, Google introduced Orchestrator in Google Antigravity. It switches between Carbon and Hydrogen within a session instead of using one model. On Terminal-Bench 4.0, Google says Orchestrator reached near-Carbon task success rates while using 24 percent less quota and finishing 19 percent faster. Teams that already use Google Gemini 4 Carbon can pair it with Hydrogen as the execution engine. For a wider view of local and cloud coding tools, the coding and developer tools directory is a good starting point.
What to watch
Hydrogen is not a frontier model. Google still positions Carbon and Argon for deep, long-horizon reasoning. Independent terminal benchmarks show Hydrogen lags behind Claude Sonnet 5.5 and GPT-6.1 Sol on agentic terminal work. Lower thinking levels are much cheaper but less intelligent: Artificial Analysis reports an index score of 41 at medium and 27 at low versus 51 at high. xhigh uses nearly twice as many output tokens as high, which explains the price jump.
For a small team shipping RAG chatbots or client products, the practical question is whether Hydrogen's speed and lower token usage justify switching from Gemini 3.8 Flash. If your workflow already uses Antigravity with Carbon, the Orchestrator combination looks worth testing. If you mostly use Search AI Mode or the Gemini app for everyday research, there is nothing to do yet. The rollout reaches free users later.
Quick answers
What is Google Gemini 4 Hydrogen?
Google Gemini 4 Hydrogen is a fast general-purpose AI model that replaces Gemini 3.8 Flash. Google released it on 12 November 2026 in Search, the Gemini app, AI Studio, the API, and Google Antigravity. It has a 1 million token context window.
How much does Google Gemini 4 Hydrogen cost?
Google lists Gemini 4 Hydrogen at $1.50 per million input tokens and $7.50 per million output tokens, the same as Gemini 3.8 Flash. Cache hits cost $0.075 per million tokens, according to The Decoder and Ars Technica.
Does Google Gemini 4 Hydrogen improve Google Search?
Yes. Google says Gemini 4 Hydrogen is now the default model in AI Mode and handles the hardest questions in AI Overviews. It understands multi-part queries and typos, and returns more visual elements and links.
Sources
- Hacker News front page: Introducing Gemini 4 Hydrogen
- The Verge: Google's new Gemini model is better at understanding complicated questions
- The Decoder: Google's Gemini 4 Hydrogen outscores the bigger Argon just four weeks after its launch
- Ars Technica: Google rolls out Gemini 4 Hydrogen, finally replacing 3.8 Flash