IndustryMulti-source

AI 使用成本每年下降 5 到 10 倍,前沿训练投入同步攀升

Why we picked it

同任务的 AI 调用成本一年降 5 到 10 倍,但前沿训练开销可能涨 10 倍。做成本预算、选模型、写商业计划的开发者都该看一眼,这两个数字能直接用进测算里。

The Information 报道称,完成同一 AI 任务的调用成本正以每年 5 到 10 倍的速度下降。Hill-Smith 同时指出,前沿实验室为训练一个能力翻倍的模型,训练开销可能扩大到原来的 10 倍。

Image · The Information
Body

这两组数字组合起来说明了行业成本结构的走向:推理端因模型优化、蒸馏和竞争性降价而快速变便宜,训练端则因追求更强能力而持续膨胀。对企业用户而言,同样任务的 API 支出在一年内可能压缩到原来的十分之一左右,模型选型策略因此需要随价格曲线动态调整。

该说法与 DeepSeek、OpenAI、Google 等近两年多轮 API 降价的公开记录方向一致,但幅度估计比多数公开降价公告更激进,可作为采购预算和成本模型的参考区间。

Original · The Information

The cost to use AI for a given task is falling 5 to 10 times a year, Hill-Smith said, even as frontier labs may spend 10 times more for a model that's twice as good.

Full story: https://t.co/65oIfHquCM

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Industry" 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 API pricing, inference cost, frontier models, The Information.

Source attribution: curated from The Information (@theinformation). Cross-source corroboration: 11 independent sources have covered this development (including Andrew Curran, rohanpaul_ai, 掘金本周最热), confirming industry-wide signal strength.