Google launched Gemini 4 Hydrogen globally on Thursday replacing Gemini 3.8 Flash across its ecosystem.
Scoring Performance on Benchmarks
Google said that Hydrogen completes tasks 2.5 times faster than Gemini 3.8 Flash, while reducing fabricated claims by more than 5x, according to Ars Technica. At its default high thinking level, Hydrogen scores 51 on the Artificial Analysis Intelligence Index, and reaches 54 at its highest xhigh thinking setting, as reported by Google.
The model offers four thinking levels—low, medium, high, and a new extra-high tier known as xhigh—which are designed to complete tasks faster with fewer output tokens, according to TechCrunch.
Google's new Gemini 4 Hydrogen has Intelligence just below Claude Sonnet 5.5 and GPT-6.1 Sol, at a third of Sonnet’s cost but ~2.5x that of GPT-6.1 Sol@GoogleDeepMind has released their third new model in six weeks. Gemini 4 Hydrogen introduces an xhigh reasoning level and… pic.twitter.com/mtn4GzkVVq
It also recorded a 71.6% score on Agents’ Last Exam v1.1, alongside a 58.8% score on HLE-Diamond.
Google DeepMind built Gemini 4 Hydrogen to tackle agentic workflows, software development, multimodal tasks, and routine coding with greater execution speed. “a major standalone upgrade for teams migrating from Gemini 3.8 Flash,” the model serves as a standalone upgrade with leaps in both speed and intelligence compared to Gemini 3.8 Flash, offering native 1 million-token input and up to 1 million output tokens.
With Gemini 4 Hydrogen live for everyone, we’ve loved seeing the “what if” ideas you’re bringing to life across @GeminiApp, @Antigravity, & @GoogleAIStudio.@LuminaBench built this 3D seaplane to fly from sunrise to night. More builds we’re loving 🧵 pic.twitter.com/t4MfHp5eql
On the Artificial Analysis Intelligence Index, Hydrogen surpasses Gemini 4 Argon at its default high thinking level, scoring 51 versus Argon’s 50, and reaches 54 at its highest xhigh thinking setting. The model features four reasoning levels and a context capacity of up to one million tokens, with Artificial Analysis assigning it an 8% hallucination rate.
Hydrogen achieves a 53.4% score on the Multimodal Knowledge Agent benchmark and 71.6% on Agents’ Last Exam v1.1 in independent benchmarks. For domain reasoning and analysis, Google states it leads on HLE-Diamond, ranks first overall on Vals Tax Agent Bench with 52.3%, ties for first on Vals Legal Research Bench with 55.8%, and leads its class on AutomationBench-AA.
Go from idea to finished video with Gemini 4 Hydrogen. Using a custom skill and simple brief, Gemini 4 Hydrogen created a sizzle reel about the element “Hydrogen," sourcing public-domain archival footage and cutting it into one continuous edit with clean narration.
Hydrogen is up to two times less likely to bypass permission boundaries when faced with impossible or blocked tasks. It achieves a 96.7% honesty rate on internal agent honesty evaluations.
Integrating Into Search and Developer Platforms
Robby Stein, VP of Product for Google Search, said that Hydrogen is used for the hardest questions in AI Overviews. The model brings improvements to Search’s fan-out technique, which breaks down single prompts into multiple subqueries to help users discover otherwise missed websites, as noted by TechCrunch.
(1/6) The biggest model upgrade to Search yet has landed. Gemini 4 Hydrogen is here, and brings cutting-edge intelligence to all Search users. It’s now the new default model in AI Mode & being used for your hardest questions in AI Overviews. > Breaks down complicated, multi-part… pic.twitter.com/XoP4jYfmao
Nick Fox of Google said that Search received a massive upgrade with the release, “biggest model improvement to Search yet,” Robby Stein, VP of Product for Google Search, described it as said.
Hydrogen enables a better understanding of user intent and provides more visual responses, including images, structured responses, comparison tables, custom tools, and additional links, bringing improvements across languages supported by AI Mode. The model handles nuanced and complicated questions more quickly and serves as the default model for AI Mode while powering harder questions in AI Overviews.
The model is available across all tiers of Google AI Studio, the Gemini API, the Gemini app, and Google Antigravity for paid Pro and Ultra subscribers, according to Google. Free users access Hydrogen in the Gemini app via automatic model selection.
The model exhibits an enhanced ability to follow and stick to tone, format, and rules while offering clear tables and visual breakdowns in the Gemini app.
🚀 @GoogleAI's Gemini 4 Hydrogen is now generally available in GitHub Copilot. Designed for fast, efficient agentic coding, tool use, and multistep workflows, it has improved instruction compared to Gemini 3.8 Flash. Find it in the GitHub Copilot app, CLI, and @code.…
Developers have access to an extra thinking tier alongside low, medium, and high levels within Google Antigravity. They also have an Orchestrator feature where Carbon can offload routine steps to Hydrogen. Running the Orchestrator with Hydrogen writing and testing the code and Carbon setting the plan attains near-Carbon results on Terminal-Bench 4.0, while operating 19% quicker on average and consuming 24% less quota, as Google pointed out.
Gemini 4 Hydrogen Costs $1.50 per Million Input Tokens
Gemini 4 Hydrogen has a flat price across its full 1 million-token context window of $1.50 per million input tokens and $7.50 per million output tokens, alongside cached input priced at $0.075 per million tokens, as reported by Memeburn and Google.
What we shipped this week @google Monday: Two new 8th-gen TPUs Tuesday: Gemini 4 Carbon – AlphaProof Code verified Signal protocol – Gemini Robotics 2.5 Wednesday: AlphaGenome 2 for disease research – Gemini Decisions API in AI Studio Thursday: Gemini 4 Hydrogen – intelligent…
“Hydrogen is built to stand on its own as a model for everyday work — or to pair with Carbon and Argon as the high-efficiency execution engine in multi-model workflows,” wrote Koray Kavukcuoglu, Google’s chief AI architect, according to Mashable.
Kavukcuoglu said that Gemini 4 Hydrogen simplifies accomplishing tasks, maintaining focus and more strictly adhering to instructions throughout multi-turn exchanges.
More on Hydrogen
Worth a look
Discover more from Archyworldys
Subscribe to get the latest posts sent to your email.