At the 2025 World Artificial Intelligence Conference (WAIC2025) held on July 26, the Shanghai Artificial Intelligence Laboratory officially released and open-sourced a scientific multimodal large model named "Shu Shen" Intern-S1. This new model aims to overcome the limitations of traditional research methods in analyzing complex phenomena, especially in deep interdisciplinary exploration.
Intern-S1 combines the strengths of the Shu Shen large model family, achieving balanced development in language and multimodal performance, and integrating rich multidisciplinary professional knowledge. As the first open-source general model that integrates professional scientific capabilities, Intern-S1 surpasses all current open-source multimodal large models in comprehensive performance, demonstrating outstanding research capabilities.
Along with the release of Intern-S1, the Shanghai AI Laboratory also launched a scientific discovery platform called Intern-Discovery. This platform aims to enhance the collaborative capabilities between researchers, research tools, and research objects, driving scientific research from team-level point exploration to larger-scale scientific discoveries. The capabilities of Intern-S1 not only enable it to surpass top closed-source models like Grok-4 in tasks across multiple fields such as chemistry, materials science, and geoscience, but also make it lead other open-source models, such as InternVL3 and Qwen2.5-VL, in multimodal comprehensive capabilities.
Intern-S1 innovatively introduced the "Cross-modal Scientific Analysis Engine," which can accurately interpret various complex scientific data, such as chemical formulas, protein structures, and seismic wave signals. This technology not only predicts the synthetic pathways of compounds but also evaluates the feasibility of chemical reactions and even identifies seismic events, providing stronger support for researchers.
In addition, Intern-S1 will continue to promote the development of a multi-agent virtual physician system, helping discover new targets and evaluate their clinical translation value. In the future, the Shanghai AI Laboratory will continue to open-source the full-chain tool system of Intern-S1, promoting broader applications and collaborations, and supporting the continuous progress of scientific research.
Intern-S1 Experience Page:
https://chat.intern-ai.org.cn/ (Click Read More at the end of the text to access directly)
GitHub Link:
https://github.com/InternLM/Intern-S1
HuggingFace Link:
https://huggingface.co/internlm/Intern-S1-FP8
ModelScope Link:
https://modelscope.cn/models/Shanghai_AI_Laboratory/Intern-S1
Key Points:
🔍 Intern-S1 is the latest multimodal large model released by the Shanghai AI Laboratory, aiming to enhance the depth and breadth of scientific research.
📈 The model's performance in fields such as chemistry, materials science, and geoscience has surpassed existing top closed-source models, demonstrating excellent research capabilities.
🤝 The Shanghai AI Laboratory will continue to promote the open-source and application of Intern-S1, supporting researchers in exploring new paths for scientific discovery together.