Google has released Gemini 4 Hydrogen, the fast, lower-cost member of its Gemini 4 family, at $1.50 per million input tokens and $7.50 per million output tokens, flat across a 1 million-token context window. Google’s own pricing page shows that rate is twice what Gemini 3.8 Flash costs today, until January 1, when Flash rises to the same price. For teams running Flash in production, the upgrade decision and the budget decision now land in the same seven weeks.
What Google announced
In a November 12 post by Koray Kavukcuoglu, Google said Hydrogen is rolling out across the Gemini API in Google AI Studio and Android Studio, Gemini for business, the Gemini app, Google Search and, for paid Google AI Pro and Ultra subscribers, Google Antigravity. Google positions it as the execution model that sits under its two frontier models, Gemini 4 Carbon and Gemini 4 Argon, and as a standalone replacement for Gemini 3.8 Flash.
Google gives four thinking levels: low, medium, high (the default) and a new xhigh tier. At high, Google says Hydrogen scores 51 on the Artificial Analysis Intelligence Index against 50 for Argon, and 54 at xhigh. Against 3.8 Flash, Google says Hydrogen finishes tasks 2.5 times faster at its default setting, scores 57% on Terminal-Bench 4.1 against 18.2%, and cuts fabricated claims by more than five times. It says partners testing graph, SQL and data-heavy workflows saw Hydrogen run 2 to 2.8 times faster while using 12% to 23% fewer tokens than 3.8 Flash.
For enterprises, Google lists three routes: Gemini for business through the API or the model menu, Google Antigravity, and Gemini in Google Sheets for eligible business subscriptions. The announcement does not mention Vertex AI or Google Cloud, so a team that buys Gemini through a Google Cloud contract should ask Google when Hydrogen reaches that contract and which data terms will cover it.
Google also introduced Orchestrator in Antigravity, which switches between Carbon and Hydrogen within one session. On Terminal-Bench 4.0, Google says Orchestrator came close to Carbon’s task success rate while using 24% less quota on average and finishing 19% faster.
How the price compares
The announcement gives Hydrogen’s price. The comparison with the rest of the line-up comes from Google’s Gemini API pricing page, which lists each model’s paid-tier rates per million tokens:
- Gemini 4 Hydrogen: $1.50 input, $7.50 output, with batch at $0.75 and $3.75.
- Gemini 3.8 Flash: $0.75 input, $3.75 output through December 31, then $1.50 and $7.50 from January 1, 2027.
- Gemini 4 Argon: $2 input, $10 output through November 30, then $4 and $20 from December 1.
Take a month of 1 billion input tokens and 200 million output tokens, the same workload AITech used when Gemini 4 Argon launched with introductory pricing ending November 30. On Hydrogen it costs $3,000. On 3.8 Flash it costs $1,500 this year and $3,000 from January. On Argon it costs $4,000 until the end of November and $8,000 from December 1.
So the two migrations look different. A team moving from Argon to Hydrogen pays 37.5% of Argon’s December rate for a model that Google says scores one point higher on the same index at default settings. A team moving from 3.8 Flash pays double the per-token price until the end of the year. If the 12% to 23% token reduction Google’s partners reported held evenly across input and output, a Flash workload would cost about 1.5 to 1.8 times as much on Hydrogen in December, and 12% to 23% less from January, when the two list prices match.
The thinking-level bill
Google’s pricing page says every token the model processes is billed, including thinking tokens, at the output rate. That makes the thinking level a cost setting as much as a quality setting. The 54 index score Google cites is at xhigh, which spends more output tokens per answer than the default. A team that benchmarks at xhigh and deploys at high should expect the score and the bill to differ from the test.
What it means for the buying committee
Three Google price events fall within seven weeks: Hydrogen arrives at a flat rate this week, Argon’s rate doubles on December 1, and 3.8 Flash’s rate doubles on January 1. A 2027 budget built on November invoices for Argon or 3.8 Flash shows half of the rate those models will charge. This continues the pattern AITech has tracked across vendors, from Anthropic’s per-token pricing for Fable 5.2 to Google’s Gemini Decisions API at $0.10 per million input tokens, where the model choice and the unit price move together.
The free tier needs a policy line. Google’s pricing page lists Hydrogen as free of charge on the AI Studio free tier, and marks free-tier content as used to improve Google’s products; on the paid tier it is marked as not used. Staff trying Hydrogen on a personal AI Studio key during the rollout would be sending company data under the free-tier terms.
What to do before January 1
Re-run your cost model with the January Flash rate. If the business case for a Flash workload only works at $0.75 per million input tokens, it has seven weeks left.
Test Hydrogen on your own traffic at the thinking level you plan to deploy. Measure tokens per completed task, since Google’s 12% to 23% reduction comes from partner tests on specific workflows.
Move eligible jobs to batch. At $0.75 and $3.75, Hydrogen’s batch rate matches the price 3.8 Flash charges today for standard requests.
For Argon workloads, list the ones where Hydrogen’s default setting is good enough before December 1, when Argon’s rate doubles.
Source: Google, Introducing Gemini 4 Hydrogen. Additional source: Google, Gemini API pricing.
