Google released Gemini 4 Hydrogen on Monday positioning the model as a faster, more efficient alternative to its frontier counterparts, Gemini 4 Carbon and Argon.
While previous releases prioritized deep reasoning, this new iteration focuses on rapid execution for coding and agentic workflows.
The model is now available across the free Gemini app, Google Search, Google AI Studio, the Gemini API, and to paid subscribers of Google AI Pro and Ultra in Google Antigravity, according to Google. By design, Hydrogen is built to handle multi-step tasks where speed is prioritized over the heavy-duty reasoning capabilities of the Carbon and Argon models.
"We saw the strongest gains in data-heavy workflows, where accurately reading tables and performing complex calculations is essential to producing reliable reports," Google stated in its official announcement. These performance improvements aim to reduce misaligned actions by 12x and increase resilience against indirect prompt injections by more than 2.3x, as measured on the Gray Swan IPI benchmark.
Turn a rough sketch into a working custom font in minutes with Gemini 4 Hydrogen. Using Google @Antigravity in the Gemini app for Mac, Hydrogen handles the technical heavy lifting so you can build custom typefaces and interactive design tools from scratch.
The release follows a staggered October 2026 rollout for the Gemini 4 family, which included the debut of Argon followed by Carbon two weeks later, Foro3D reported. While Carbon and Argon remain the go-to for complex, long-horizon problems, Hydrogen completes tasks 2.5 times faster than the previous Gemini 3.8 Flash model, TechCrunch noted.
In domain-specific evaluations, Hydrogen leads its class on the HLE-Diamond benchmark with a 58.8% score and ranks first on the Vals Tax Agent Bench at 52.3%. It also outperformed the more powerful Gemini 4 Argon on the Artificial Intelligence Index at default high-thinking levels, scoring 51 compared to Argon’s 50, according to TechCrunch.

For Google Search, the model introduces an improved "fan-out" technique that breaks user prompts into multiple subqueries. This allows the search engine to provide more visual responses, including comparison tables and structured data, specifically when handling nuanced or multi-part queries.
Google has implemented a flat pricing structure for Hydrogen, charging $1.50 per million input tokens and $7.50 per million output tokens across its 1 million-token context window. While the model is integrated into free tiers, certain advanced features within Google Sheets remain exclusive to paid Google AI Pro or Ultra subscribers, blog.google confirmed.
What happens when you put Gemini 4 Carbon to work in the lab? 🔬 Researchers are using our latest model in Google @Antigravity to tackle multi-step problems across physics, math, and the life sciences. pic.twitter.com/pe1QLnFxh8
Earlier internal testing of the model, which engineers described as "very capable, fast and efficient," occurred on an internal platform referred to as Jetski, Business Insider reported. Despite the public launch, some details regarding the long-term free availability of specific Hydrogen-powered tools remain unconfirmed, as Google declined to provide further comment on its internal strategy.
A little visual ASMR to kick off your week 💧 Created with Gemini 4 Hydrogen in @GeminiApp. Hydrogen is available to everyone, so you can bring your ideas to life across @GeminiApp, Google Search, @GoogleAIStudio, and more. pic.twitter.com/gJY6ii02Gs
Worth a look