Large Language Model Series
7.2k 2026-04-30
InternLM/InternLM
A series of high-performance, cost-effective open-source large language models (LLMs) designed for general-purpose usage and advanced reasoning.
Core Features
State-of-the-art performance on reasoning and knowledge-intensive tasks.
Significantly reduced training costs (e.g., 75% less for InternLM3-8B) while maintaining leading performance.
Supports deep thinking mode for complex reasoning and normal response mode for fluent interactions.
Offers a comprehensive model zoo including InternLM, InternLM2, InternLM2.5, and InternLM3 series with various sizes.
Open-sourced instruction models for broad applicability.
Detailed Introduction
InternLM is an official release of a series of large language models, including InternLM, InternLM2, InternLM2.5, and InternLM3. These models are characterized by their enhanced performance in reasoning and knowledge-intensive tasks, often surpassing competitors like Llama3.1-8B and Qwen2.5-7B. A key innovation is their cost-efficiency, achieving top-tier results with significantly less training data. InternLM models also feature a unique deep thinking capability for complex problem-solving alongside fluent interaction modes, making them versatile tools for various AI applications.