原帖引用 · AI 翻译
标题:“焦耳到 GDP:AI 系统每单位能源能产生多少经济价值?”
斯坦福的一篇新论文发现,较小的 AI 模型每单位能源所创造的经济价值可能远高于能力最强的模型。
如今世界受到能源和电力的限制,确实应该研究每项 AI 任务消耗多少能源。
在这里,研究人员构建了 Joules-to-GDP,用 AI 工作的美元价值减去其运行成本,再除以所消耗的焦耳数。他们用 GDPval 中涵盖 44 种职业的 220 项专业任务,测试了 5 个开放模型。
> GLM 5.2 每项任务产生的净价值最高,但 Gemma 4 31B 仅消耗其 23% 的能源,就创造了其 56% 的净价值。
> 将 Qwen3.6-27B 使用的 H100 数量从 1 个增加到 8 个,只将任务耗时缩短了 22%,却消耗了 3.2 倍的能源。
> 尽管价格和速度与 GLM 5.2 几乎相同,GPT-OSS 120B 消耗的能源却少了 68%。
> 同一模型使用 8 块 H100 芯片而非 1 块时,任务完成速度仅提高 22%,能源消耗却达到 3.24 倍,而所完成工作的美元价值几乎相同
https://t.co/KAj3vxhHto
原文
New Stanford paper finds that a smaller AI model can return far more economic value per unit of energy than the most capable one. Now that the world is limited by energy/power, how much energy each AI task burns should indeed by studied. Here, the researchers built Joules-to-GDP, which measures the dollar value of an AI's work, minus its running cost, per joule of energy used. They tested 5 open models on 220 professional tasks from GDPval, covering 44 occupations. > GLM 5.2 produced the most net value per task, but Gemma 4 31B got 56% of it with 23% of the energy. > Moving Qwen3.6-27B from 1 to 8 H100s cut task time by just 22% while using 3.2x the energy. > GPT-OSS 120B used 68% less energy than GLM 5.2 despite nearly the same price and speed. > Running the same model on 8 H100 chips instead of 1 made tasks only 22% faster but used 3.24x the energy, for nearly the same dollar value of work https://t.co/KAj3vxhHto
AI 消耗的全球能源越来越多。但我们从 AI 获得了多少经济价值?
我们推出 Joules-to-GDP(J2G),这是一个衡量 AI 将能源转化为经济价值的效率,并显示哪些任务、模型和硬件带来最佳回报的框架。
基于 @OpenAI 的 GDPval 中金融、工程、法律、行政工作等领域的 44 种职业,我们得出了 3 项发现:
1)经济价值最高的任务,并非每焦耳经济价值最高的任务。例如,CAD 设计每项任务的收益比建网站高约 10%,但后者每焦耳创造的价值高约 27%。
2)最强大的模型不一定是最有效的能源利用方式。Gemma 4 31B 仅消耗 GLM 5.2 的 23% 的能源,就实现了其 56% 的净价值,每焦耳创造的价值达到其 2.5 倍。
3)增加芯片数量并不能保证回报同比增加。将同一模型使用的 NVIDIA H100 数量从 1 个增加到 8 个,能源消耗达到 3.2 倍,但完成时间只缩短 22%,创造的价值大致相同。
我们还与 @SnorkelAI 合作发布 GDPWorks,这是一个涵盖另外 36 种职业中具有经济价值的任务的基准测试。
这项合作研究由 @amirzeinaalii 和 @Avanika15 共同牵头,参与者还有 @Mersad_Abbasi、@Jon_Hartley_、John Hennessy、@HazyResearch 和 @Azaliamirh。
更多详情见下文 🧵
原文
AI is using more and more of the world’s energy. But how much economic value are we getting from AI? We’re introducing Joules-to-GDP (J2G), a framework that measures how well AI turns energy into economic value and shows which tasks, models, and hardware give the best return. Across 44 occupations in finance, engineering, law, administrative work, and more in @OpenAI's GDPval, we arrive at three findings: 1) The tasks with the highest economic value are not the tasks with the highest economic value per joule. For example, CAD design earns ~10% more per task than building websites, but the latter earns ~27% more value per joule. 2) The most powerful model isn’t always the best use of energy. Gemma 4 31B delivers 56% of GLM 5.2’s net value with 23% of the energy, for 2.5x the value per joule. 3) More chips don’t guarantee proportionate returns. Moving the same model from 1 to 8 NVIDIA H100s uses 3.2x the energy but cuts completion time by only 22%, with about the same value. With @SnorkelAI, we’re also releasing GDPWorks, a benchmark for economically valuable tasks across 36 additional occupations. Joint work co-led by @amirzeinaalii and @Avanika15, with @Mersad_Abbasi, @Jon_Hartley_, John Hennessy, @HazyResearch, and @Azaliamirh. More details below 🧵