轨迹分析揭示Gemini 4 Carbon编程运行较慢的原因
AI 总结
一项私有软件工程评测显示,Gemini 4 Carbon有时在困难缺陷修复任务上耗时接近一小时,而Claude Opus 5.5和GPT-6 Astra仅用数分钟。轨迹分析将差距归因于推理速度较慢、运行轨迹更长、不必要的代码审计、重复读取文件以及每步工具调用更少。在分析的运行中,Carbon的步骤数约为Astra的七倍,动作数约为十一倍。展示的单次试验中,Carbon低档位耗时22分21秒,而Opus 5.5高档位耗时9分33秒。
深度见解
该分析指出了与Carbon在困难编程智能体任务中延迟较高相关的具体推理、行为和机械效率因素,包括不必要的工作和重复读取文件。
Why is Gemini 4 Carbon taking so long?
November 2026
While benchmarking Gemini 4 Carbon on one of our private evals, we found something surprising: it spent close to an hour completing tasks that routinely took Opus and Astra only a few minutes:
Time and cost per trial
Left is faster; lower is cheaper.
API list price On a $200 plan, fully used
#Gemini 4 Carbon#Claude Opus 5.5#GPT-6 Astra#Gemini 4 Hydrogen#opencode#mini-SWE-agent