Skip to content

InternThinker

A reasoning model with self-reflection and correction

Visit official site

What InternThinker is

The model is described as a strong-reasoning AI model from Shanghai AI Laboratory that reflects on and corrects its own reasoning path.

What you can do with it

  • Generate high-density data for complex reasoning
  • Reflect on and correct a reasoning path
  • Build a high-quality chain of thought independently

Getting started

  • Log in to the trial platform
  • Select the InternThinker model
  • Enter a complex reasoning task
  • Watch the shown reasoning process and meta actions

Pros & cons

✓ What we like

  • Chains of thought are built without existing strong-reasoning models
  • Sandbox feedback supports self-learning and evolution

! What to watch out for

  • Reasoning runs through long chains of thought
  • Feedback comes from a large-scale sandbox environment

FAQ

Who developed it?

The material says it was developed by Shanghai AI Laboratory.

What is meta-action thinking?

The material describes self-reflection and correction with steps such as understanding, recall, planning and execution.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review