Google releases Gemini 4 Hydrogen, its new fast model

Gemini 4 Hydrogen is Google's fast, cheaper model: $1.50 per million input tokens, a 1M context, and now inside Search.
What it means for founders
If you run Gemini 3.8 Flash in production, Hydrogen is the obvious test. The price is flat across the whole context window, so long-document jobs are easier to budget. Rerun your own evaluations at each thinking level before switching; Google's benchmark wins are on its chosen tests, and Google's token savings (12% to 23% in partner tests) may not hold on your workload.
The bigger pattern is the pairing: a strong model plans, a cheap one executes. Google is building that into Antigravity, and teams can copy it in their own agents to cut cost per task.
Search matters too. If AI Mode fans out across more sub-queries, pages that answer one narrow part of a question well have more chances to be linked. Watch your Search Console referrals over the next few weeks.
What we don't know yet
Google has not said when Gemini 3.8 Flash will be retired from the API. Independent benchmark results and real-world latency under load are not out yet. It is also unclear how much the Search changes will shift traffic to publishers, which Google did not quantify.
See Enki's page on Gemini for plans and pricing.
The story
Google has released Gemini 4 Hydrogen, the fast, cheaper member of the Gemini 4 family. Google pitches it as an upgrade over Gemini 3.8 Flash, which remains available, and it is rolling out globally on the same day to the Gemini API, every AI Studio tier, the Gemini app, Google Search and Gemini for business.
Carbon and Argon, the frontier models Google released earlier (Enki covered the Carbon launch), are pitched at long, hard reasoning. Hydrogen is pitched at the work in between: writing code, running tests, research and data tasks that need speed more than depth.
The numbers Google is claiming
Input costs $1.50 and output $7.50 per million tokens, the same rate across the whole 1 million token context window. Developers pick from four levels of thinking, from low to a new xhigh, with high as the default.
Against Gemini 3.8 Flash, Google says Hydrogen is 2.5 times as fast on its default setting, scores 57% on Terminal-Bench 4.1 against Flash's 18.2%, and makes fabricated claims more than five times less often. It scores 51 on Artificial Analysis's Intelligence Index at high, one point above Argon, and 54 at xhigh. These are Google's figures; independent tests will follow.
Google also announced Orchestrator in Google Antigravity, which switches between Carbon and Hydrogen within one session. The company says that on Terminal-Bench 4.0 it came close to Carbon's success rate while using 24% less quota and finishing 19% faster on average.
Search gets it too
Every Google Search user worldwide can now get Hydrogen, starting with its AI answers. Google calls it the "biggest model upgrade to Search yet" and says it handles multi-part questions, typos and very specific requests better. It also widens the "fan-out" step that splits a query into sub-searches, which Google says surfaces a broader range of links, alongside more tables and visual answers.
Sources
Primary sources
Tools in this story
We may earn a commission if you sign up through our links. It never affects our ratings or which stories we cover.
Google's assistant, built into Workspace and Android
More in Models & Labs
- Nvidia's open source push faces test after Huang exit

For founders: Hosting terms. The pledge that Hugging Face stays chip neutral is Nvidia's, made in a post under Huang's name.
TechCrunch · 5d ago - Google releases Gemini 4 Carbon, previews /swarm

For founders: Budget for the December price. Cost models built on the $2 and $10 rate break after November 30.
Google DeepMind · 8d ago - Anthropic releases Fable 5.2 and Mythos 5.2

For founders: Price the switch by task cost. At twice the Opus 5.5 rate, Fable needs to finish jobs with fewer tokens or retries to save money.
The Verge · 8d ago - Meta launches Muse Ember for high-volume agent work

For founders: Price your agent by the task. Ember's advantage depends on short outputs and cheap cache hits, so a pipeline that reuses a long system prompt gains the most.
TechCrunch · 9d ago - OpenAI's GPT-6.1 Terra halves per-token API rates

For founders: Rerun your agent costs. At $1 in and $5 out per million tokens, high-volume chores like invoice entry and form filling get cheaper.
OpenAI · 23d ago