Shanghai AI Laboratory develops the InternLM series of large language models.

InternLM is a family of open LLMs (InternLM, InternLM2, InternLM2.5, InternLM3) supporting context windows up to 1 million tokens.

Code and weights are published on GitHub at InternLM/InternLM and on Hugging Face under the internlm organization.

InternLM2.5 released in July 2024, with InternLM3 following in early 2025.

Created to give researchers and enterprises a free, commercially usable, high-context Chinese/English LLM.

Deployable via Hugging Face Transformers, LMDeploy, or vLLM for inference.

Code under Apache 2.0, weights open for research and free commercial use, no credentials needed, self-hosted; maintained by Shanghai AI Laboratory.