LLM Research Papers: A Curated List for July to December 2025

Summary

Sebastian Raschka presents a curated list of large language model research papers he bookmarked from July through December 2025. The list is based mainly on skimming abstracts, and he notes that he read only a small fraction of the papers in full. It is organized into sections covering reasoning-model training, inference-time reasoning and evaluation, other reinforcement-learning methods, inference-time scaling, model releases and technical reports, architectures, efficient training, diffusion-based language models, multimodal and vision-language models, and data and pre-training datasets. Raschka originally planned to include the collection in his annual LLM review, but separated it into its own post because that review had become lengthy. He presents the collection as an organized reference that he can revisit for future projects, rather than as a comprehensive survey or detailed analysis of every paper.