InternAI (Shusheng)

InternAI (Shusheng) is a pioneering general large model system developed by the Shanghai AI Laboratory. It represents a family of advanced, open-source foundation models designed to push the boundaries of artificial intelligence across various scientific and industrial domains. The system is not a single tool but a comprehensive ecosystem of specialized models, each tailored for specific, complex tasks.

How to use InternAI (Shusheng)

As a suite of foundation models, InternAI (Shusheng) is primarily aimed at developers, researchers, and enterprises. Usage methods vary depending on the specific model:

• Open-Source Models: Models like Shusheng·Puyu (LLM) and Shusheng·Wanxiang (Multimodal) are open-sourced. Developers can download the model weights and source code from platforms like Hugging Face or GitHub to integrate them into their own applications, conduct research, or fine-tune them for specific tasks.
• API Access: For certain applications and enterprise-level integrations, the models may be accessible via APIs. This allows for leveraging the models' power without the need for managing the underlying infrastructure. Users would typically need to register on the official platform or contact the laboratory for an API key.
• Dedicated Platforms: Specialized models like Shusheng·Fengwu (Weather) or Shusheng·Yifei (Aerospace) may be offered through dedicated web platforms or systems where users can interact with the model's capabilities directly for forecasting or design tasks.

Core Features of InternAI (Shusheng)

• Shusheng·Puyu (LLM): An advanced large language model with leading reasoning capabilities, a 1 million token long context window for deep document understanding, and the ability to autonomously integrate information from over 100 web pages to complete complex tasks.
• Shusheng·Wanxiang (Multimodal): A world-leading open-source multimodal model whose performance is comparable to closed-source commercial models. It supports a wide range of modalities and is highly efficient, reducing training costs by up to 80%.
• Shusheng·Fengwu (Weather): A revolutionary AI-powered weather and climate forecasting platform. It provides precise short-term precipitation forecasts and has extended the effective global medium-range forecast period to over 10 days.
• Shusheng·Yifei (Aerospace): The world's first generative AI system for airfoil design. It allows for controllable, explainable, and intelligent generation of supercritical airfoils, embedding professional constraints for next-generation aircraft design.
• Shusheng·Tianji (3D Modeling): A city-scale 3D reality model based on NeRF, capable of large-scale urban reconstruction and editing with high efficiency in training and real-time rendering.
• Shusheng·Jishi (Finance): A specialized large model for the financial sector, enhanced with retrieval augmentation for accurate, up-to-date financial analysis, reasoning, and compliance checks.
• Shusheng·Yanpu (Science): The first large model applied to analyze Electron Paramagnetic Resonance (EPR) spectra, achieving expert-level accuracy and reducing analysis time from hours to seconds.

Use Cases for InternAI (Shusheng)

The diverse models within the InternAI (Shusheng) family cater to a wide array of applications:

• Scientific Research: Accelerating discoveries in chemistry, materials science, and biology with Shusheng·Yanpu; advancing climate science with Shusheng·Fengwu.
• Engineering & Design: Optimizing aircraft design with Shusheng·Yifei; creating digital twins and urban planning simulations with Shusheng·Tianji.
• Finance: Automating financial report analysis, risk management, and investment decision support with Shusheng·Jishi.
• Content Creation: Generating high-quality text, code, and creative multimodal content with Shusheng·Puyu and Shusheng·Wanxiang.
• Information Services: Powering complex search engines, intelligent assistants, and long-document summarization tools.

Advantages of InternAI (Shusheng)

The primary advantage of the InternAI (Shusheng) system is its comprehensive and open approach. It provides a robust, open-source foundation that lowers the barrier to entry for high-performance AI development. By offering a family of specialized, state-of-the-art models, it enables innovation in critical sectors that require domain-specific knowledge and high accuracy. Its focus on efficiency, such as reducing training costs and enabling deployment on consumer-grade hardware, makes advanced AI more accessible to a broader community of developers and researchers.

Pricing and Plans

Many of the models within the InternAI (Shusheng) ecosystem, such as Shusheng·Puyu and Shusheng·Wanxiang, are open-source and available for free for research and commercial use (subject to specific license agreements). For dedicated platform access, enterprise solutions, or customized deployments, interested parties are typically required to contact the Shanghai AI Laboratory directly for partnership and pricing information. This hybrid model supports both the open-source community and specialized industrial applications.