Gemini 4 Hydrogen (High) 登上 Agent Arena 第 5 名，主打速度与成本
Google DeepMind 新模型 Gemini 4 Hydrogen 上线，Agent Arena 第 5 名，速度是 Claude Fable 5.2 的 3 倍，成本只有 GPT-6 Astra 的零头。跑智能体工作流、对 API 账单敏感的团队可以直接看下价格对比。
Google DeepMind 发布 Gemini 4 Hydrogen (High)，在 Agent Arena 上首次亮相即排名第 5，基于 6.5k 条真实智能体会话，净改进幅度达 +14.8%。它的核心卖点是速度和成本：中位数 5.3 分钟完成任务，是前十名中最快的模型，比第 1 名 Claude Fable 5.2 快 3 倍，比 GPT-6 Astra 快 2.4 倍；成本为每任务中位数 1.6 美元，比 GPT-6 Astra 便宜 3.7 倍，比 Claude Fable 5.1 便宜 4.7 倍。API 定价为每百万 tokens 输入 1.5 美元、输出 7.5 美元。
技术上，Hydrogen 每任务消耗更多输出 tokens：比 Carbon 多 41%，是 Argon 的 2.8 倍，说明它倾向于用更长的推理换速度和成本优势。关键信号上，可操控性（Steerability）排第 2（+21%），仅次于 Fable 5.2 的 +33%，高于 Carbon 的 +19% 和 Opus 5.5 的 +18%；Praise vs. Complaint 排第 4（+32%），Confirmed Success 排第 5（+15%）。
分任务类别看，Chat 排第 3（+19%）、Work 排第 7（+13%）、Code 排第 9（+12%），官方定位为快速编程、多模态合成和日常工作流。官方还称其任务完成速度比 3.8 Flash 快 2.5 倍，并在 Artificial Analysis Intelligence Index 上超过 Argon。对需要高频调用智能体的开发者来说，它在性能第一梯队里提供了明显的价格折扣选项。
Gemini 4 Hydrogen (High) by @GoogleDeepMind debuts #5 on Agent Arena with +14.8% net improvement across 6.5k real-world agentic sessions. Hydrogen stands out for its speed and cost efficiency: - Fastest model in the top 10: 5.3 min/task (median), 3x faster than #1 Claude Fable 5.2 and 2.4x faster than GPT-6 Astra - Cost-efficient: $1.6 cost/task (median), 3.7x cheaper than GPT-6 Astra and 4.7x cheaper than Claude Fable 5.1. API token pricing: $1.5/$7.5 input/output per million tokens Notably, it spends more output tokens per task: 41% more than Carbon, 2.8x more than Argon. Strong across key signals: - #2 Steerability: +21% (behind only Fable 5.2 at +33%; ahead of Carbon at +19% and Opus 5.5 at +18%) - #4 Praise vs. Complaint: +32% - #5 Confirmed Success: +15% Ranking by task category: - #3 in Chat: +19% - #7 in Work: +13% - #9 in Code: +12% Hydrogen delivers competitive performance while completing tasks significantly faster and at a fraction of the cost of several leading models. Congrats to the @GoogleDeepMind team on another impressive release! Google DeepMind @GoogleDeepMind Meet Gemini 4 Hydrogen: built for rapid coding, multimodal synthesis, and everyday workflows. It finishes tasks 2.5x faster than 3.8 Flash and surpasses Argon on the @ArtificialAnlys Intelligence Index. 🔗 View Quoted Tweet📊 6 ⚡
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 DeepMind, Agent Arena, Claude Fable.
Source attribution: curated from lmarena.ai (@lmarena_ai). Cross-source corroboration: 11 independent sources have covered this development (including Artificial Analysis, berryxia, Philipp Schmid), confirming industry-wide signal strength.
