Skip to content

CogVLM2

A bilingual multimodal dialogue model with 8K context

Visit official site

What CogVLM2 is

The model is described as a second-generation multimodal pretrained dialogue model built by a Tsinghua University team.

What you can do with it

  • Work with benchmarks such as TextVQA and DocVQA
  • Use the Chinese and English open-source versions

Pros & cons

✓ What we like

  • Bilingual support is listed
  • Basic calling methods and fine-tuning examples are provided

! What to watch out for

  • The material compares it with the previous generation
  • Open-source versions cover Chinese and English

FAQ

Which resolutions are listed?

The material lists 1344*1344 images and 8K content length.

Are fine-tuning examples included?

The material lists basic calling methods and fine-tuning examples.

Last reviewed: 2026-11-15

More AI chatbot tools

View all →

How we review