DEV Community

Cover image for How to Connect MiniMax-M3 to Cursor
Orchid Files
Orchid Files

Posted on Edited on

How to Connect MiniMax-M3 to Cursor

#cursor #minimax #ai #opensource

The Missing Thinking and Reasoning Separation in Cursor

The documentation on the MiniMax website states that the model uses the OpenAI format and is compatible with most IDEs. But it turned out that if you connect it to Cursor, it doesn't separate the contents of <think>...</think> from the actual response; everything goes into a single stream. You see the model's thoughts and the response mixed together, and at some point it becomes unclear where its thoughts end and the response to the user begins. Working in this mode is extremely inconvenient.

 

How Ungate Fixes It

I've added MiniMax support to the Ungate extension for Cursor that I'm developing. It processes content from <think>...</think> blocks, and now Cursor correctly separates the model's reasoning from the response to the user. Working with MiniMax in Cursor is now the same as with other OpenAI-compatible models.

Without Ungate, the model's thoughts and the answer arrive in one stream — a single text block where the reasoning and the response run together:

<think>... I need to check the project structure first. Let me list the files
and then propose a plan...</think> But the actual answer continues right here,
so it is unclear where the thinking ends and the response begins.
Enter fullscreen mode Exit fullscreen mode

With Ungate, Cursor correctly separates the model's reasoning from the response, and only the answer reaches the chat:

Here is the final plan for the task.
Enter fullscreen mode Exit fullscreen mode

Login and tunnel setup for the extension are the same for MiniMax as for other models — see How to use Claude and ChatGPT subscriptions in Cursor.

 

Connecting MiniMax in Cursor

To access the model, you need to add the custom model name to Cursor:

MiniMax-M3
Enter fullscreen mode Exit fullscreen mode

I've added a Base URL selector for MiniMax to the Ungate settings:

  • China
  • Global
  • Custom

 

Quality and Speed: Benchmarks vs Practice

MiniMax-M3 is a new Chinese frontier model from MiniMax. According to some benchmarks, it has almost caught up with Claude Opus 5. However, based on my tests over the past few days, I've concluded that it doesn't even measure up to Claude Sonnet 5. If you use it for simple tasks, everything is fine.

But if you have a monorepo project structure with packages and apps, you have to run a lot of iterations to complete tasks. Even if the rules and skills specify the project structure — where types, helpers, and ESLint configurations are located — it still doesn't follow that structure. And it's very slow compared to Sonnet, and it's about 20 times slower than GPT-5.6. I asked the agent with MiniMax to copy the structure from another monorepo repository, and it took me 4 hours of back-and-forth with the agent to clarify the details so that it would finally complete the task. With Claude Sonnet 5, this takes me about 15–30 minutes.

Benchmarks and real-world performance differ greatly. If a model is estimated to be close to Sonnet or even Opus in benchmarks, in practice there may be a significant gap between them.

 

Pricing and Value

You need to consider not only quality and speed, but also price. MiniMax is 10 times cheaper than Claude Sonnet 5. A $10 subscription gives you:

  • 1,500 queries every 5 hours
  • 15,000 queries per week

I also noticed that MiniMax's planning capabilities are quite good — comparable to Sonnet or GPT. Therefore, for simple tasks or situations where speed isn't a priority, MiniMax can be a reasonable choice.

 

My Blog: orchidfiles.com
Telegram Channel: @orchidfiles

Top comments (0)