Scientific reasoning in materials science requires integrating multimodal experimental evidence with underlying physical theory. Existing benchmarks make it difficult to assess whether incorporating visual experimental data during post-training improves mechanism-grounded explanation reasoning beyond text-only supervision. We introduce MATRIX, a multimodal benchmark for materials science reasoning that evaluates foundational theory, research-level reasoning, and the interpretation of real experimental artifacts across multiple characterization modalities. Using MATRIX as a controlled diagnostic, we isolate the effect of visual grounding by comparing post-training on structured materials science text alone with post-training that incorporates paired experimental images. Despite using relatively small amounts of multimodal data, visual supervision improves experimental interpretation by 10-25% and yields 5-16% gains on text-only scientific reasoning tasks. Our results demonstrate that these improvements rely on correct image-text alignment during post-training, highlighting cross-modal representational transfer. We also observe consistent improvements on ScienceQA and PubMedQA, demonstrating that the benefits of structured multimodal post-training extend beyond materials science. The MATRIX dataset is available at https://huggingface.co/datasets/radical-ai/MATRIX and the model at https://huggingface.co/radical-ai/MATRIX-PT.
Multimodal Fine-grained Reasoning for Post Quality Evaluation
School of Computer Science(计算机科学学院);School of Artificial Intelligence Institute(人工智能研究所);State Key Laboratory of Information Security, Institute of Information Engineering(信息安全国家重点实验室,信息工程研究所);Key Laboratory of Education Informatization for Nationalities(民族教育信息化重点实验室)
A longitudinal geospatial multimodal dataset of post-discharge frailty, physiology, mobility, and neighborhoods
University of Toronto(多伦多大学);KITE Research Institute, Toronto Rehabilitation Institute, University Health Network(大学健康网络多伦多康复研究所KITE研究院);American University of the Middle East(中东美国大学)
Beyond Atomic Geometry Representations in Materials Science: A Human-in-the-Loop Multimodal Framework
Computer Engineering, Texas A\&M University, College Station, TX 77843, USA;College of Science;Engineering, Hamad Bin Khalifa University, Doha, Qatar;Dept. of Electrical \& Computer Engineering, Texas A\&M University at Qatar, Doha, Qatar;Orthotics, Ankara University, Ankara, Turkey
Understanding the Capabilities of Molecular Graph Neural Networks in Materials Science Through Multimodal Learning and Physical Context Encoding
Dept. of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77843, USA(电气与计算机工程系,德克萨斯A&M大学);College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar(科学与工程学院,哈马德·本·卡西姆大学);Dept. of Prosthetics and Orthotics, Ankara University, Ankara, Turkey(假肢与矫形学系,安卡拉大学)
Physics-Guided Multimodal Transformers are the Necessary Foundation for the Next Generation of Meteorological Science
School of Artificial Intelligence, Beijing University of Posts;Huawei Noah's Ark Lab;Department of Earth System Science, Tsinghua University;College of Computing \& Data Science, Nanyang Technological University;State Key Laboratory of Networking;Switching Technology, Beijing University of Posts
Earth Science Foundation Models: From Perception to Reasoning and Discovery
Department of Data Science and Artificial Intelligence, The Hong Kong Polytechnic University(数据科学与人工智能系,香港理工大学);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室)
Understanding the Role of Hallucination in Reinforcement Post-Training of Multimodal Reasoning Models
University of North Carolina at Chapel Hill(北卡罗来纳大学教堂山分校);University of Science and Technology of China(中国科学技术大学);Arizona State University(亚利桑那州立大学);Honda Research Institute, USA(本田美国研究所)
Maven: A Multimodal Foundation Model for Supernova Science
The NSF AI Institute for Artificial Intelligence and Fundamental Interactions(美国国家科学基金会人工智能与基础相互作用AI研究所);Department of Physics, Harvard University(哈佛大学物理系);Institute for Advanced Computational Science, Stony Brook University(石溪大学高级计算科学研究所);Massachusetts Institute of Technology(麻省理工学院);Center for Astrophysics — Harvard & Smithsonian(哈佛-史密森天体物理中心);Center for Theoretical Physics, Massachusetts Institute of Technology(麻省理工学院理论物理中心);Anthropic(Anthropic公司)
MoDoMoDo: Multi-Domain Data Mixtures for Multimodal LLM Reinforcement Learning
Brown University(布朗大学);Salesforce AI Research(Salesforce AI研究院);NVIDIA Research(英伟达研究院);Carnegie Mellon University(卡内基梅隆大学);Princeton University(普林斯顿大学);Texas A&M University(德克萨斯A&M大学);California Institute of Technology(加州理工学院);MIT CSAIL(麻省理工学院计算机科学与人工智能实验室)