LLM Research Papers: List for the Second Half of 2025
Sebastian Raschka presents a curated selection of interesting research papers on large language models from July to December 2025. The material includes categories such as reasoning models, training methods, architectures, multimodal models, and more. The list is intended for quick reference and further study.
In June 2025, Sebastian Raschka published a bonus article for paid subscribers featuring a selection of research papers. Now, as a thank-you to all supporters, he has prepared a similar list for July–December 2025, organized by category: reasoning models (training, inference-time strategies, evaluation and understanding), other reinforcement learning methods, inference-time scaling, model releases and technical reports, architectures, efficient training, diffusion language models, multimodal and vision-language models, and data and pretraining sets. The author himself briefly skimmed the abstracts and read only a small portion, but he finds such lists useful to revisit when working on projects. It is noted that the collection was originally intended as part of the annual review "State of LLMs 2025: Progress, Challenges, and Predictions," but the article became too long, so Raschka decided to publish the list separately.
Source: Sebastian Raschka —
original