Agent Arena 分析：Gemini 4 Carbon 在高负载任务上成本优势扩大
Arena.ai 用自家 Agent Arena 数据对比了几款旗舰模型的成本曲线，发现 Gemini 4 Carbon 任务越难越划算：重度会话里 GPT-6 Astra 相对成本翻倍，Opus 5.5 到 4.2 倍。常跑长时程智能体任务、在意 API 账单的开发者可以看看选型账怎么算。
Arena.ai 的 Agent Arena 分析显示，Google DeepMind 的 Gemini 4 Carbon 在任务难度提升时成本竞争力反而增强。典型任务下，GPT-6 Astra 的成本是 Carbon 的 3.4 倍，Claude Opus 5.5 为 3.9 倍，GPT-6.1 Sol 约为 Carbon 的一半（0.54 倍）。
在成本最高的 5% 会话中，Carbon 相对所有对手的优势进一步扩大：Opus 的成本倍数升至 4.2 倍，Astra 相对 Carbon 的成本大约翻倍，而 Sol 相对 Carbon 的价格折扣则收窄。这意味着 Carbon 在长时程任务上的开销随任务复杂度增长得更慢。
文中以构建 Android 安全应用这类长时程任务为例，说明成本差异会转化为实际可观的节省。对于大量使用智能体执行复杂任务的用户，选择 Carbon 可能在重度场景下明显降低支出，而轻量场景下 GPT-6.1 Sol 仍是更便宜的选择。
Our recent Agent Arena analysis found that Gemini 4 Carbon by @GoogleDeepMind becomes more cost-competitive as tasks get more demanding. Its savings relative to GPT-6 Astra and Claude Opus 5.5 grow on the most expensive sessions. Arena.ai @arena Models by @GoogleDeepMind scale better on cost than their rivals. Gemini 4 Carbon is the prime example. As tasks get more demanding, every rival gets more expensive relative to Carbon. For a typical task, GPT-6 Astra by @OpenAI costs 3.4× as much, Claude Opus 5.5 by @AnthropicAI costs 3.9× as much, and GPT-6.1 Sol is the cheaper option at about half of Carbon’s cost (0.54×). Across the 5% most expensive sessions, Carbon’s position improves against all three. Opus edges up to 4.2×, Astra’s cost roughly doubles relative to Carbon, and Sol’s discount shrinks. This in turn can translate into meaningful savings across real-world tasks such as building an Android security app, one of many long-horizon tasks that our Agent Arena users delegate to their agents.
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 Carbon, GPT-6 Astra, Claude Opus 5.5, 推理模型.
Source attribution: curated from lmarena.ai (@lmarena_ai).
