Skip to content

InternLM-XComposer-2.5

A vision-language model with long context

Visit official site

What InternLM-XComposer-2.5 is

The model is a versatile large vision-language model with long context input and output on a 7B LLM back end.

What you can do with it

  • Understand ultra-high-resolution images
  • Handle fine-grained video understanding
  • Hold multi-turn multi-image dialogue

Creation tasks

  • Write page code in HTML, CSS and JavaScript
  • Produce high-quality illustrated articles
  • Extend context to 96K through RoPE extrapolation

Pros & cons

✓ What we like

  • Writes HTML, CSS and JavaScript page code
  • Supports multi-turn multi-image dialogue

! What to watch out for

  • A 7B LLM back end is used
  • Environment setup is required

FAQ

How long is the context?

The material states 24K training context and up to 96K.

What can it write?

The feature list names page code and illustrated articles.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review