Gemini 4 Hydrogen Is Now on Multos — The Model Google Tells 3.8 Flash Users to Move To
Google's own migration guide points Gemini 3.8 Flash users at Hydrogen. Same 1M unified context, a fourth thinking level, and five times fewer fabricated claims. It is live on Multos Pro today at $1.50/$7.50.
Google calls Gemini 4 Hydrogen the recommended upgrade from Gemini 3.8 Flash, and for once that framing is not marketing. It runs $1.50 per million input and $7.50 per million output, flat across the full 1M context window, and it cuts fabricated claims by more than 5x compared to 3.8 Flash. It is on Multos Pro and above as of today.
The Spread of Things You Can Build
Hydrogen is a general-purpose workhorse, which means it fits a wide range of builds rather than one niche. A multi-tenant SaaS dashboard with auth, billing, and a database. A customer portal that handles requests and status. An internal tool that reads a spreadsheet export and turns it into something a team can act on. A content pipeline that drafts, checks, and publishes. A data-entry app with validation. A booking flow. A reporting tool. What these have in common is that they are ordinary software builds where you want the model to be competent and cheap enough that iterating does not feel expensive. Hydrogen is priced to be the default you reach for, not the special occasion.
Built for Parallel Work, Not Solo Chat
The capability that separates Hydrogen from a plain chat model is how it behaves when you give it several jobs at once. Google positions it for parallel worker and verifier agents: one agent writes, another checks, a third handles a different slice of the task, all running against the same 1M window. If you have built anything with subagents on Multos, you know the failure mode is a worker that drifts or a verifier that rubber-stamps. Hydrogen was trained for that pattern specifically, and it batches read and search tool calls in parallel rather than walking them one at a time. That is a latency difference you feel on any build with more than a couple of files.
A Fourth Thinking Level
Hydrogen ships with four thinking levels: low, medium, high as the default, and a new xhigh. This is more useful than it sounds for app building. A form validation fix does not need deep reasoning and should not pay for it. A schema migration across a live database does. Being able to set the level per task, rather than picking a different model, means one build can span both without you re-architecting anything. Google's own guidance is to start at high and reserve xhigh for hard multi-step problems, because xhigh increases latency and token use and can overthink simple prompts.
The Specs, Plainly
One unified 1,048,576-token window covering input, thinking, and output together. Default synchronous output is 65,536 tokens, extendable to 1M with long decode. Input accepts text, images, video, audio, and PDF. Function calling, structured outputs, code execution, and computer use are all supported. Context caching is implicit only, which means cache hits depend on a stable prompt prefix rather than an explicit cache you manage. If your build sends a large system prompt on every turn, keep it byte-stable and the discount applies on its own.
What It Costs on Multos
Hydrogen is a Pro model, available on Pro and Elite. At $1.50 input, $7.50 output, and $0.075 cached, the blend works out to $0.664 per million at our standard cache assumption, which is a 1.33x token multiplier. That puts it slightly below Gemini 4 Carbon at $0.885 and well below the Claude Opus tier. For a model Google is positioning as the everyday default, that is the right shape: cheap enough to leave selected, capable enough that you are not switching constantly.
Should You Move Off 3.8 Flash
If you are on 3.8 Flash today, the case is straightforward. Hydrogen is the model Google says to move to, it costs more per token but fabricates far less, and the 1M unified window is larger than what you are used to. The tradeoff is tier: 3.8 Flash is available on Starter, Hydrogen starts at Pro. If you are on Starter and 3.8 Flash is working for your build, there is no urgency. If you are on Pro or Elite and doing anything where a confidently wrong answer costs you a debugging session, the move pays for itself quickly.
Ready to Get Started?
Gemini 4 Hydrogen is live on Multos Pro and above. 1M unified context, four thinking levels, and five times fewer fabricated claims than 3.8 Flash, at $0.664 per million blended. Open the model dropdown, find it under Google, and make it your default for the next thing you build.
Start Building FreeRelated Articles
One-Shot Website Generation: Why Most AI Builders Can't Do It (and How Multos Does)
Most AI builders need 10-30 prompts to reach a finished website — and every prompt costs credits. Multos generates a complete, designed, deployed site in one shot. Here is the engineering reason one-shot works, and why competitors structurally can't match it.
Read more →GPT-5.6 Luna Price Cut 80% — Now Available on Free & Lite
OpenAI slashed GPT-5.6 Luna by 80% overnight — from $1/$6 to $0.20/$1.20 per million tokens. We're passing it straight through: Luna is now available on every plan including Free and Lite.
Read more →Claude Opus 5 & Sonnet 5 Now Available: Near-Frontier Intelligence at Every Tier
Anthropic's Claude Opus 5 and Claude Sonnet 5 are now live on Multos. Opus 5 delivers near-Fable-5 intelligence at half the price with 79.2% on SWE-bench Pro, while Sonnet 5 brings Opus-class agentic capabilities to Starter-tier users at $3/$15 per million tokens.
Read more →