Google 发布 Gemini 4 Hydrogen：对标 Claude Sonnet 5.5 与 GPT-6.1 Terra
Google 这次把推理、速度和 1M 上下文塞进一个模型，比 3.8 Flash 快 2.5 倍，还用 Carbon 编排器调度多模型。如果你在 Claude、GPT 和 Gemini 之间挑主力模型，这条值得对比着看，做长视频或大文件分析的开发者尤其该关注。
Google 发布 Gemini 4 Hydrogen，官方称其推理能力达到 Claude Sonnet 5.
5 与 GPT-6.1 Terra 的水平。
速度方面，该模型比上一代 3.8 Flash 快约 2.
5 倍。模型支持 1M token 上下文窗口，并原生支持视频与文件输入。
同时，Antigravity 产品中加入了动态多模型编排器 Carbon，可调度多个模型协同工作。这一版本将长上下文、多模态输入与速度提升打包在同一模型中，直接与 Anthropic 和 OpenAI 的旗舰推理模型竞争。
Gemini 4 Hydrogen is one of our coolest model drops.
• Claude Sonnet 5.5 & GPT-6.1 Terra level reasoning • super fast to work with (~2.5x faster than 3.8 Flash) • 1M context with native video & file support • dynamic multi-model orchestrator in Antigravity with Carbon https://t.co/t1x9cYYtBV
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 US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Gemini 4 Hydrogen, Google, Claude Sonnet 5.5, GPT-6.1 Terra.
Source attribution: curated from Shubham Saboo (@Saboo_Shubham_). Cross-source corroboration: 11 independent sources have covered this development (including 宝玉, Blognone, IT之家), confirming industry-wide signal strength.
