InternLM/Intern-S1

A Scientific Multimodal Foundation Model

What it solves

Intern-S series provides a collection of multimodal foundation models designed specifically for scientific intelligence and long-horizon agentic tasks. It addresses the gap in open-source models' ability to handle complex scientific reasoning, interpret specialized modalities (like molecular formulas and protein sequences), and execute multi-step scientific workflows.

How it works

The series employs several advanced training and architectural strategies across different model versions:

  • Visual Pre-training: Intern-S2 models learn directly from raw scientific literature pages, modeling symbolic semantics and visual relationships together to preserve text-visual correspondence without needing intermediate parsing.
  • Task Scaling & RL: The models use large-scale multi-task reinforcement learning (RL) across 20+ scientific domains. Intern-S2-Preview-35B specifically focuses on "task scaling" to unlock capabilities in a smaller parameter footprint.
  • MoE Architecture: High-capacity versions like Intern-S1-Pro use a trillion-scale Mixture-of-Experts (MoE) architecture with 512 experts to balance performance and efficiency.
  • Specialized Modalities: The models utilize dynamic tokenizers and Fourier Position Encoding (FoPE) to natively understand chemical structures, protein sequences, and long heterogeneous time-series signals.
  • Agentic Training: Models are connected to sandboxed environments for black-box reinforcement learning to improve their ability to handle long-horizon tasks.

Who it’s for

  • Scientific Researchers: Those needing AI assistants for biomolecular interaction design, material structure generation, or compound synthesis planning.
  • AI Developers: Engineers building long-horizon agents for complex, multi-step workflows.
  • ML Researchers: Those studying multimodal reasoning and the scaling of scientific intelligence.

Highlights

  • Scientific Specialization: Strong performance in chemistry, materials science, and life sciences.
  • Multimodal Capability: Native understanding of raw scientific pages, molecular formulas, and seismic signals.
  • Diverse Model Sizes: Ranges from lightweight (Intern-S1-mini) to trillion-scale (Intern-S1-Pro).
  • Efficiency Gains: Intern-S2-Preview-35B uses MTP and CoT compression to improve token generation speed and reasoning efficiency.

Related

  • Project
    OpenGVLab/InternVideo

    A series of video foundation models and large-scale datasets designed for multimodal video understanding, long-context modeling, and contextual reasoning.

  • Project
    MMMU-Benchmark/MMMU

    A massive multi-discipline benchmark for evaluating multimodal AI models on college-level subject knowledge and complex reasoning across 30 different subjects.

  • Project
    SkyworkAI/Skywork-R1V

    Skywork-R1V is an open-source multimodal reasoning model that uses reinforcement learning and visual chain-of-thought to achieve state-of-the-art performance in complex visual logic, math, and physics tasks.

  • Project
    microsoft/unilm

    A comprehensive collection of foundation models and architectures from Microsoft focusing on large-scale self-supervised pre-training across text, vision, speech, and multimodal tasks.

  • Project
    zai-org/GLM-4

    A series of open-source LLMs featuring 9B and 32B parameter models with specialized variants for general chat, deep reasoning, and complex rumination tasks.