S1-Omni is trained on S1-Omni-Corpus, which covers 200 scientific tasks and contains millions of reasoning samples, and is evaluated on over 60 scientific benchmarks, providing a practical path toward unified scientific modeling.
Abstract
We present S1-Omni, a unified multimodal reasoning model for scientific understanding, prediction, and generation. AI for Science (AI4S) has advanced significantly through domain-specific models, tool-augmented LLMs, and scientific language models. However, model capabilities remain highly fragmented, limiting the joint modeling of heterogeneous data, scientific laws, and expert knowledge. S1-Omni addresses this gap by consolidating these capabilities into a single, coherent scientific reasoning model. The architecture of S1-Omni is built upon three core components: unified representation of scientific data, natural-world knowledge alignment, and decoding for domain-specific tasks. First, S1-Omni maps natural-language instructions and scientific objects, including CIF, SMILES, protein sequences, spectra, and scientific images, into a shared representation space. Second, it incorporates scientific laws and expert knowledge into data construction and training, enabling the model to reason from scientific evidence. Third, it performs task-specific decoding to support a broad range of applications, including property prediction, spectrum-to-molecular generation, protein site and structure prediction, and scientific image generation and editing. S1-Omni is trained on S1-Omni-Corpus, which covers 200 scientific tasks and contains millions of reasoning samples, and is evaluated on over 60 scientific benchmarks. It outperforms GPT-5.5 and Gemini-3.1-Pro on most benchmarks and matches or surpasses domain-specific models on several benchmarks. Overall, S1-Omni provides a practical path toward unified scientific modeling.
Experiments show that MKB achieves competitive scientific understanding across biological and molecular benchmarks, produces high-fidelity native outputs for weather forecasting, biological generation, and medical-image segmentation, and largely retains the general capabilities of its Qwen3-VL backbone.
Hesen Chen, Xinyue Su, Xiaomeng Yang et al.· 0 citations
Scientific discovery increasingly requires AI systems that can reason over scientific evidence of heterogeneous modalities, interact with scientific tools and environments, and sustain progress across long task horizons. We present Intern-S2-Preview, a series of scientific agentic foundation models designed to support multimodal scientific understanding, reasoning, generation, and long-horizon tasks. The training pipeline begins with scientific multimodal pre-training over rendered scientific documents, interleaved image-text data, and diverse scientific corpora. Starting from the pretrained checkpoint, we apply a unified post-training pipeline consisting of supervised fine-tuning, scalable multi-task reinforcement learning (RL), black- and white-box agentic RL, and on-policy distillation. This pipeline is supported by practical techniques that improve rollout and training stability and efficiency, including partial rollout with off-policy correction, adaptive length regularization, online speculative decoding, robust multi-task optimization, and trace-aware experience assembly for agentic tasks. At the architecture level, Intern-S2-Preview-397B extends time series modelling from efficient long-sequence understanding to numerical forecasting, while Memory Decoder is studied as a separate memory-augmented path for rapid scientific specialization without modifying the frozen 397B backbone. Evaluations across scientific, multimodal, agentic, and general-purpose benchmarks show that Intern-S2-Preview-397B achieves competitive or leading results in multiple settings. The time series modules improve scientific signal understanding and forecasting on SciTS, while the separate Intern-MemDec-4B extension improves the Biology-Instructions average score from 56.92 to 60.32 without modifying the frozen 397B backbone.
Driven by multimodal large language models (MLLMs), remote sensing image (RSI) understanding is undergoing a paradigm shift, evolving from learning a domain-specific model to learning a general foundation model with domain adaptation (LaGD). Under the LaGD paradigm, conventional datasets, such as DOTA and RSICD, which fueled progress in RSI understanding over the past decade, are no longer adequate for emerging tasks because their annotation formats are task-specific and lack the language-level supervision required by MLLMs. We argue that a new dataset must be purposefully designed to support three core capabilities: First, generalization, enabling models to learn shared knowledge across tasks through a unified annotation format; Second, complex scene understanding, training models to capture fine-grained object attributes and spatial relationships and to describe scenes in detailed natural language; Finally, reasoning, equipping models with high-level visual reasoning through multiturn dialogues. To this end, we present UniRS-Instruct, a high-quality, diversified, and unified multimodal instruction-following dataset for RSI understanding. UniRS-Instruct unifies diverse tasks, including image captioning, visual question answering, visual grounding, and region-level captioning, into a consistent (question and answer) format. To construct fine-grained and context-aware instruction data, we propose a hierarchical prompting strategy: at the local level, objects are identified via rotated bounding boxes to describe their fine-grained attributes and spatial relationships; at the global level, local information is integrated with the full image to generate detailed scene-level instruction descriptions through GPT-4 V. Extensive experiments on multiple remote sensing benchmarks demonstrate that MLLMs fine-tuned with UniRS-Instruct achieve superior performance in image captioning, visual question answering, and visual grounding tasks, and exhibit stronger capabilities in describing fine-grained information, uncovering implicit knowledge, and conducting complex reasoning compared with models trained on existing datasets.
Linrui Xu, Yuhan Wang, Ling Zhao et al.· IEEE Journal of Selected Top...· 0 citations
Multimodal Large Language Models (MLLMs) have demonstrated strong abilities in solving diverse visual and textual reasoning tasks. However, their development in the physics domain is significantly hindered by the lack of a comprehensive benchmark. To fill this gap, we introduce OmniPhys, a large-scale benchmark for multimodal physics understanding and reasoning, covering middle school through university-level problems from Chinese Educational Corpora. OmniPhys consists of 15,246 questions and 19,850 images, accompanied by detailed annotations that support fine-grained analysis of reasoning processes and knowledge usage. Beyond conventional evaluation, OmniPhys is a benchmark that systematically evaluates multimodal outputs in the physics domain, including models'ability to generate structured physics diagrams, which constitute a fundamental component of authentic physics problem solving. Extensive evaluations reveal critical gaps in the capabilities of current MLLMs, especially in complex reasoning and visual generation. To address this, we release OmniPhys to serve as a foundational resource for advancing multimodal intelligence in physics and scientific domains. Codes and data are available at https://github.com/ECNU-RAIL/OmniPhys-EMNLP2026.
Hao Chen, Yu-Min Lin, Nadila Yushanjiang et al.· 0 citations
This survey reviews the evolution of language models from early statistical approaches to modern Transformer-based architectures and summarizes key developments, including attention mechanisms, scaling laws, alignment techniques, and efficient inference methods.
P. Peykani, V. Charles, Ali Emrouznejad et al.· Archives of Computational Me...· 0 citations
Multimodal Large Language Models have made great progress in grounding tasks, yet existing methods still struggle to unify precise localization and complex reasoning. For one thing, text-based methods rely on coordinates or index prediction, severely limiting the perceptual capabilities of the model for dense visual objects. Meanwhile, latent token-based methods employ special tokens without inherent spatial references and use a decoding mechanism that lacks thinking steps, weakening high-level reasoning capabilities. Consequently, developing a unified framework that excels in both perception and reasoning remains challenging. To address this, we propose Mixture-of-Thought-Tokens (Motto), a new free-form multimodal grounding method that bridges the perception-reasoning gap, enabling MLLMs to empower diverse, arbitrary grounding queries. Specifically, we introduce Spatially-Grounded Thought Tokenization to explicitly align special tokens with spatial locations for clear spatial correspondence and visual interpretability. We further design a Context-Adaptive Chain-of-Tokens that dynamically switch grounding modes within an interleaved reasoning chain, achieving robust grounding across tasks of varying complexity. In addition, we construct PR-Bench, a new referring expression comprehension benchmark to evaluate the perception-reasoning gap. Extensive experiments demonstrate that Motto achieves state-of-the-art performance across diverse free-form grounding tasks.
Tianyi Gao, Han Fang, Tianyi Ding et al.· 0 citations