aiqbench analysis
This study quantifies the efficiency of large language models by measuring their economic output relative to energy consumption, revealing that smaller or differently configured models can often deliver superior value per joule compared to larger counterparts. The research highlights critical hardware dependencies and operational trade-offs, demonstrating that scaling GPU clusters does not linearly improve efficiency and that specific model architectures vary significantly in their ability to perform tasks like software development versus engineering under energy constraints.
Joules-to-GDP: Measuring the economic value of AI per unit of energy
The model that produces the most valuable work isn’t always the best use of your energy. This is one of the interesting findings in Joules-to-GDP, a paper released today by researchers at Stanford, UC Berkeley and UT Austin. Of the five models they tested, GLM 5.2 delivered the most net value per task, but Gemma 4 31B delivered 56% of... The post Joules-to-GDP: Measuring the economic value of AI per unit of energy appeared first on Snorkel AI.
The model that produces the most valuable work isn’t always the best use of your energy. This is one of the interesting findings in Joules-to-GDP, a paper released today by researchers at Stanford, UC Berkeley and UT Austin. Of the five models they tested, GLM 5.2 delivered the most net value per task, but Gemma 4 31B delivered 56% of that value on 23% of the energy, about 2.5 times the net value per joule. Our research team provided expert baselines for the 220 GDPval tasks used in the study and partnered to develop GDPWorks, 83 new tasks covering 36 occupations that GDPval doesn’t include. The most valuable model isn’t the most efficient one. Gemma 4 31B produced 44% less output value than GLM 5.2, but it returned the most net value per unit of energy of the five models. (Net value is the dollar value of the AI’s work minus what it cost to run.) Energy shows differences that cost and speed hide. GPT-OSS 120B and GLM 5.2 had nearly the same…
Excerpt shown under the publisher's RSS terms · 173 words · full text and figures are on the publisher's site