What Google released on 12 November
On 12 November 2026, Google introduced Gemini 4 Hydrogen.
The post calls it a faster model for everyday coding and agent work.
Google says Gemini 4 Carbon and Gemini 4 Argon are for longer, harder reasoning.
The same page says it was last updated on 13 November 2026.
What a free Gemini app can use
Google’s help page is called About Gemini models.
It says the table is for a personal Google account.
A work or school account follows that organization’s Workspace license.
A personal account with no Google AI plan can use Auto or Flash-Lite in chat.
With auto on, most prompts use Flash-Lite.
A prompt that needs more reasoning may use Hydrogen.
With auto off, Google says every reply uses Flash-Lite.
The help page does not list Hydrogen as a named choice on a free account.
People with a Google AI plan can switch models by hand.
Plus, Pro and Ultra can select Hydrogen for chat.
Pro and Ultra can also use Hydrogen for tasks.
The free row lists no task model.
Google says the table can change without notice.
The blog says free app users get Hydrogen only through auto model selection.
Google says the rollout is global.
Neither page names the Philippines.
What the API costs
Google prices the API at $1.50 per million input tokens.
Output costs $7.50 per million tokens.
The price is flat across a 1 million token window.
The model can take 1 million input tokens.
Google says output can run up to 1 million tokens.
A token is a small piece of text the price counts.
The post does not give a peso price.
Where else Google is rolling it out
Google says the rollout started on 12 November and is global.
Developers can use Hydrogen in the Gemini API and in Google AI Studio.
Google says AI Studio includes every tier.
The post also names Android Studio.
Paid Google AI Pro and Ultra subscribers can try it in Google Antigravity.
Businesses can deploy it in Gemini for business.
In Google Sheets, Google starts with Pro and Ultra subscribers.
Eligible business subscriptions are included there too.
Search is a separate surface.
Google says Hydrogen is reaching Search for all users in that global rollout.
The help page above is about the Gemini app, not about Search.
What Google claims about speed
Google says Hydrogen has four thinking levels.
They are low, medium, high and xhigh.
High is the default.
At the default high setting, Google says Hydrogen is two and a half times faster than Gemini 3.8 Flash.
Google cites Terminal-Bench 4.1 against Gemini 3.8 Flash.
Google says that score rises from 18.2 percent to 57 percent.
Google says Hydrogen scores 51 on the Artificial Analysis Intelligence Index at the high setting.
Google says Argon scores 50 on that same setting.
At xhigh, Google says the Hydrogen score is 54.
The Gemini app help page lists Low, Medium and High.
It does not name xhigh.
These speed and index figures are Google’s own tests.
Why it matters in the Philippines
Many students and workers here use the free Gemini app and Google Search.
On a free personal account, most prompts stay on Flash-Lite.
Hydrogen is a possible auto result, not the model on every answer.
After a reply, open the details and read the model name.
On a computer, Google says the model is at the bottom of the thinking steps.
On the phone apps, Google says the model is in the response details.
Try one question you can check, on a task you can undo.
Developers who use the API can compare $1.50 and $7.50 with the job before they switch.
What to do next
- Open Google’s 12 November 2026 post and the About Gemini models help page.
- On a free personal account, leave auto selection on if a harder prompt should have a chance of Hydrogen.
- If you turn auto off, Google says the replies stay on Flash-Lite.
- After an answer, open the reply details and read the model name.
- If you pay for the API, check $1.50 and $7.50 in AI Studio before you switch a project.
- Treat the speed and index figures as Google’s own tests.
How this article checked the pages
The release date, the API price and the product list come from Google’s post of 12 November 2026.
That page says it was last updated on 13 November 2026.
The free-account table comes from Google’s help page, About Gemini models.
This check opened the blog on 15 November 2026.
It also opened the computer, Android and iPhone views of that help page the same day.
Those three views listed the same free-account rule.
The notes on students and workers in the Philippines are this article’s reading of who may care.
Google does not say the model is already on every account in the Philippines.
The speed and index figures stay Google’s.
Put this into practice
Before you pay for the API, or trust a speed figure, check Google’s pages and try one small task you can undo.
Sources & further reading
Source note. The linked sources support the news in this article. The explanations and suggested next steps are from Pinoy AI Works.


