Gemini 4 Hydrogen in Iran
The hard-working, low-cost member of the Gemini 4 family, on Hormouz since its release day. Made for high request volume, paid in rial, with no monthly subscription.
What is Hydrogen made for?
Google released Gemini 4 Hydrogen on November 12, 2026: a fourth-generation model tuned for speed and cost rather than the very hardest problems. It sits between Gemini 3.8 Flash and the Argon and Carbon flagships, and fits chatbots, customer support, summaries, data extraction, and agents that are called thousands of times a day.
Hydrogen always thinks before it answers, and you set how deeply with four levels: low, medium, high (the default) and xhigh; it cannot be switched off completely. It reads text, images, PDFs, audio and video, and fits about a million tokens in one request.
Hydrogen next to its siblings
See price, context and capabilities number by number, then choose.
What does a month on Hydrogen cost?
Three high-volume workloads at live Hormouz prices. Thinking tokens are counted as output, so the real figure also depends on the thinking level.
| Model | Summarising 50,000 documents a month | Classifying 200,000 customer messages | A support agent handling 8,000 conversations |
|---|---|---|---|
| Gemini 4 Hydrogen | $413 | $165 | $105 |
| Gemini 3.8 Flash | $227 | $90.75 | $57.75 |
| Gemini 4 Carbon | $550 | $220 | $140 |
| GPT-6 Luna | $27.50 | $11.00 | $7.00 |
Persian is light on tokens with Hydrogen
Tokens each model spends on 100 words of one fixed Persian text; fewer means cheaper. Measured by Hormouz through the API.
Set up in a few minutes
- 1Open a Hormouz account with your mobile number.
- 2Create an API key in the panel and point your OpenAI SDK base URL at Hormouz.
- 3Set the model to google/gemini-4-hydrogen and tune the thinking level with reasoning_effort.
Hand your high-volume work to Hydrogen.
Start with the signup gift credit, then pay only for usage.
Create an accountFAQ
- How much does Gemini 4 Hydrogen cost?
- $1.50 per million input tokens and $7.50 per million output tokens; repeated input that Google caches automatically is $0.075. There is no launch discount, so this price does not change after November.
- Hydrogen or Gemini 3.8 Flash?
- Hydrogen is the newer generation, and Google presents it as the upgrade from 3.8 Flash. If answer quality matters more to you than a few cents, try Hydrogen; the comparison page puts both prices side by side.
- Hydrogen or Carbon?
- Carbon is for the hardest coding and reasoning, and costs more. Hydrogen is for high volume and quick answers. Many teams send everyday work to Hydrogen and the hard problems to Carbon.
- Can I turn thinking off?
- No. Hydrogen always thinks; reasoning_effort set to low gives the lowest cost and the quickest answer. The values none, minimal and max are not accepted.
- Does it work from Iran and with OpenAI code?
- Yes. No VPN, Google account or foreign card needed, using the OpenAI SDK with Hormouz's base URL. Google's own SDK and Gemini CLI do not connect.
Prices updated: 2026-11-16