Google 发布 Gemini 4 Hydrogen，改进复杂问题理解能力
Google’s new Gemini model is better at understanding complicated questions
Google 把 Gemini 4 Hydrogen 直接塞进了搜索框，能拆解带好几个子问题的提问，连打错的错别字也能看懂。和只答一半的旧搜索 AI 模式相比，明显更聪明，经常用搜索的人今天就能试到。
Google 发布新模型 Gemini 4 Hydrogen，主打处理复杂提问，包括含多个子问题、拼写错误明显或高度具体化的请求。该模型能够将问题拆解成多个部分，并在回答中逐一覆盖。发布当日即上线，覆盖 Google 搜索、Gemini 应用以及所有付费层级。
从机制上看，Gemini 4 Hydrogen 的关键在于问题分解：面对一个混合了多个意图的提问，模型会先识别出各个子问题，再在最终回答里分别回应，而不是只抓住其中一个意图作答。对拼写错误的容忍度提升，也意味着用户不需要反复修正输入就能拿到可用结果。
这次发布把新模型直接接入搜索入口而非单独的产品页面，说明 Google 在把 Gemini 系列的能力逐步下沉到每日高频的搜索场景中，与 OpenAI 把模型放进 ChatGPT 主入口的路径类似。对普通用户来说，变化体现在搜索 AI 模式回答长尾、复合问题时更完整。
Google’s new Gemini model is better at understanding complicated questions
Google says its new Gemini 4 Hydrogen model can give more helpful answers to complex questions, including "questions with multiple parts, hard-to-understand typos, or highly specific requests." The new model can break down your questions and address each part in its responses. It's available starting today in search, the Gemini app, all tiers of Google […]
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Gemini 4 Hydrogen, Google, 搜索, 多模态.
Source attribution: curated from theverge (@Stevie Bonifield). Cross-source corroboration: 11 independent sources have covered this development (including lmarena.ai, Logan Kilpatrick, Google Gemini App), confirming industry-wide signal strength.
