Tracking account reports pricing in the billing catalog
A post saying that pricing for "Gemini 4 Hydrogen," an unannounced Google model, had appeared in Google Cloud's billing catalog went up on X on Nov. 12 (5:33:54 p.m. Nov. 12 Japan time). According to @aitrackerbot, an account that describes itself as automatically tracking AI model releases, the price per million tokens is $1.50 for input (text, image, audio and video), $0.075 for input using caching (a discount for repeated input) and $7.50 for output. It said "Batch" and "Flex" processing cost half that, while "Priority" processing costs $2.70 for input and $13.50 for output.
Gemini 4 Hydrogen pricing is now in Google Cloud's billing catalog 💸 $1.50 per 1M input tokens (text, image, audio or video) $0.075 cached input $7.50 output Batch and Flex are half that, Priority is $2.70 / $13.50. Under half of Gemini 4 Carbon's $4 / $20.
Hydrogen is a code name, and since late October posts by individuals have said it corresponds to Gemini 4 Flash, but Google has not confirmed this. On Nov. 5, ORIENS reported claims that it was being A/B tested on AI Studio and on the comparison site Arena, and this is a follow-up to that report. Google has made no announcement, and ORIENS has not been able to confirm the prices on publicly available official pages.
Leaker's image says release is "unconfirmed"
About two minutes earlier, at 5:31:42 p.m. the same day, the leaker @LuminaBench posted that Hydrogen was in the billing catalog and also wrote that "It's finally the day for release."
Gemini 4 Hydrogen is in the Google Cloud billing catalogue It's finally the day for release https://t.co/iC2oxUZXtz
The attached image is a notification screen marked as developed by @LuminaBench. It names the service as "Vertex AI" and lists standard processing at $1.5 for input and $0.075 for cache reads, and Batch and Flex at $0.75 for input, $3.75 for output and $0.0375 for cache. It cites 69 billing items (SKUs) as evidence, while also stating "Billing listing only; release, public model ID and API access unconfirmed." The image does not show an output price for standard processing.
At 5:36:30 p.m. the same day, @LuminaBench followed up with a post saying Hydrogen's pricing will be $0.75 for input, $3.75 for output and $0.0375 for cache, and wrote that this was "Not bad at all with how good it is." These figures are the same as the Batch and Flex prices in its own image, and the post does not say which type of processing they apply to. It also offered no material showing the model's performance.
Gemini 4 Hydrogen's pricing will be $0.75 Input $3.75 Output $0.0375 cache Not bad at all with how good it is https://t.co/3JZIg8PwxG
Name absent from official pricing pages and model list
The Gemini API pricing page and model list, and the Google Cloud Vertex AI pricing page, which ORIENS retrieved in the 6 p.m. hour on Nov. 12, list neither the Hydrogen name nor its prices. The Gemini 4 models in the model list are the top-end "Gemini 4 Carbon" and the efficiency-focused "Gemini 4 Argon."
Google Cloud's public search page for billing catalog items did not include the list of items in the content ORIENS retrieved, so ORIENS could not confirm whether Hydrogen is listed. For now, the $1.50 input and $7.50 output prices remain reports from the two accounts. Google has also not given a release date.
Reported price is less than half of Carbon's price from December
@aitrackerbot compared the price with Carbon's $4 for input and $20 for output, saying it was less than half. According to the Gemini API pricing page, Carbon has an introductory price tied to its general availability of $2 for input and $10 for output through Nov. 30, and $4 for input and $20 for output is the price from Dec. 1. Compared with Carbon from December, the reported price is $1.50 versus $4 for input and $7.50 versus $20 for output. Compared with the introductory price through Nov. 30, it is $1.50 versus $2 for input and $7.50 versus $10 for output.
The current lightweight model "Gemini 3.8 Flash" is listed on the same pricing page at $0.75 for input, $3.75 for output and $0.075 for cache through Dec. 31. From Jan. 1, 2027, it rises to $1.50 for input and $7.50 for output. The Hydrogen price reported by @aitrackerbot is at the same level as 3.8 Flash's price from 2027.
Addendum: On Nov. 13, Google officially released the model (Google: Introducing Gemini 4 Hydrogen).
- Gemini 4 Hydrogen (reported)
- Gemini 3.8 Flash (through Dec. 31)
- Gemini 4 Carbon (from Dec. 1)
