aiqbench analysis
This study reveals a significant divergence in the AI landscape, where model builders are rapidly increasing parameter counts while scientific adoption lags behind, potentially limiting research impact. The data indicates that vision models currently dominate usage in science, though language model integration is accelerating, with open-weight variants leading the way. Researchers may be missing out on higher-impact publication opportunities by not utilizing larger foundation models available today.
Science Is Falling Behind the Frontier: Foundation Model Adoption Across Half a Million Papers
arXiv:2511.21739v2 Announce Type: replace-cross Abstract: We present the first large-scale analysis of AI foundation model usage in science -- not just citations or keywords. We find that adoption has grown rapidly, at nearly-exponential rates, with the highest uptake in Linguistics, Computer Science, and Engineering. Vision models are the most used foundation models in science, although language models' share is growing. Open-weight models dominate. As AI builders increase the parameter counts of their models, scientists have followed suit but at a much slower rate: in 2015, the mean foundation model adopted in science was 5.4x larger than the mean model being built; by 2024 that relationship had reversed, with the mean model built 6.9x larger than the mean model adopted. We also present suggestive evidence that scientists' use of these smaller models may be limiting them from getting the full benefits of AI-enabled science, as papers that use larger models appear in higher-impact journals and accrue more citations.
Abstract:We present the first large-scale analysis of AI foundation model usage in science -- not just citations or keywords. We find that adoption has grown rapidly, at nearly-exponential rates, with the highest uptake in Linguistics, Computer Science, and Engineering. Vision models are the most used foundation models in science, although language models' share is growing. Open-weight models dominate. As AI builders increase the parameter counts of their models, scientists have followed suit but at a much slower rate: in 2015, the mean foundation model adopted in science was 5.4x larger than the mean model being built; by 2024 that relationship had reversed, with the mean model built 6.9x larger than the mean model adopted. We also present suggestive evidence that scientists' use of these smaller models may be limiting them from getting the full benefits of AI-enabled science, as papers that use larger models appear in higher-impact journals and accrue more citations.
22 pages (8 main text), 6 figures, 3 tables. Accepted to the AI for Meta-Science (AI4MetaScience) Workshop at NeurIPS 2026
Digital Libraries (cs.DL); Artificial Intelligence (cs.AI)
arXiv:2511.21739 [cs.DL]
(or arXiv:2511.21739v2 [cs.DL] for this version)
https://doi.org/10.48550/arXiv.2511.21739 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 267 words
