zoom_in Click to enlarge

InternLM2.5

▲ ▼

language

description InternLM2.5 Overview

InternLM2.5 is an open-weight large language model series developed by Shanghai AI Laboratory and released in 2024 as part of their continued InternLM lineup. The series includes variants with different parameter sizes and a flagship model capable of processing up to 1 million tokens of context. InternLM2.5 demonstrates strong performance on benchmarks for reasoning, coding, and mathematical tasks, with particular strengths in Chinese language understanding. The models are released with commercial-friendly licenses to support both research and industrial applications.

help InternLM2.5 FAQ

How large is InternLM2.5's context window?

InternLM2.5 includes a variant with a 1-million-token context window, which was among the largest available in any open-weight model at the time of its 2024 release. This massive context enables processing extremely long documents, code repositories, or extended conversations in a single prompt.

Who developed InternLM2.5?

InternLM2.5 was developed by Shanghai AI Laboratory, one of China's prominent AI research institutions. The InternLM series is part of the lab's broader open-source AI initiatives, which also include frameworks like OpenCompass for model evaluation.

Is InternLM2.5 open source?

Shanghai AI Laboratory released InternLM2.5 model weights openly, making them available through Hugging Face and their own GitHub repository. The releases include multiple model sizes to support different deployment scenarios and research needs.

What improvements does InternLM2.5 offer over the original InternLM?

InternLM2.5, released in 2024, improved over the first-generation InternLM series in areas including mathematical reasoning, coding proficiency, and tool use. The introduction of the million-token context window variant was one of the headline upgrades, along with enhanced data quality and training techniques.

swap_horiz

Looking for InternLM2.5 alternatives? Compare top competitors ranked & scored

arrow_forward