Skip to content

Kimi视觉思考模型k1

Reason over images with chain-of-thought output

Visit official site

What Kimi视觉思考模型k1 is

The model is described as an AI model built with reinforcement learning that natively supports end-to-end image understanding and chain-of-thought reasoning. Its capabilities extend beyond mathematics into basic science fields, and the material reports benchmark ability tests in mathematics, physics and chemistry.

What you can do with it

  • Process image input directly without external OCR or vision models
  • Form a detailed chain of thought through reinforcement learning
  • Work across mathematics, physics and chemistry questions
  • Handle image and graphic information with optimised character recognition

Example uses

  • A student solves a complex geometry figure problem
  • A teacher shows the steps of a physics circuit question in class
  • A researcher examines a chemistry technical diagram

Getting started

  • Get the latest mobile assistant app or open the web version at kimi.com
  • Find the visual thinking entry on the Kimi+ page and take or upload an image
  • Read the chain of thought shown as the model works toward an answer
  • Swipe to view the full chain of thought and long-press to download it

Pros & cons

✓ What we like

  • The reasoning chain is shown while the model works toward an answer
  • Character recognition is optimised for image and graphic information

! What to watch out for

  • Access starts from the mobile app or the web version at kimi.com
  • An image has to be taken or uploaded before reasoning begins

FAQ

What is Science Vista?

The feature list says the Kimi team built a standardised test set of mathematics, physics and chemistry image questions at varied difficulty.

How do I see the reasoning chain?

The documented flow shows the chain of thought on screen, with swiping for the full chain and a long press to download it.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review