MiniMax M2.1: Coding Model, Pricing and Limits
minimax/minimax-m2.1MiniMax M2.1 is a text model for coding, reasoning, tool use, and agent-style workflows.
205K tokens of context — about 153,600 words, on a scale that ends at 2M.
What it can do
- Reasoning
- Tool use
- Open weights
- Takes text
- Gives text
Released 23 Dec 2025 · checked 3 Oct 2026.
About it
MiniMax M2.1 is a text-only language model from MiniMax, released December 23, 2025. It supports reasoning and tools, focuses on coding and multi-step workflows, and has open weights. The model accepts up to 204,800 tokens at once and can produce up to 131,072 tokens in one answer. Usage is priced separately for input and output tokens.
Review
In a line
MiniMax M2.1 is a text model made by MiniMax. It is designed for coding, reasoning, tool use, and workflows that involve several steps.
What it's good at
The model is aimed at coding and modern application development. Its reasoning support suits tasks that need the model to work through a problem rather than only give a short reply.
Tool support lets it work with external actions when the surrounding software provides them. It handles text input and produces text output, not images. See other models with tool calling.
Price in plain terms
Input costs $0.3 per million tokens. Output costs $1.2 per million tokens.
Reading a page of about 500 words costs about 0.019 cents. A chat turn with about 1,000 input tokens and 500 output tokens costs about 0.09 cents. At 1,000 such turns, the rough total is about $0.90.
Context and limits
M2.1 can take up to 204,800 tokens in one request. That is the amount of text and other input the model can consider at once.
One answer can contain up to 131,072 tokens. The model is text-only, so it does not accept or create images.
Ways to use it
The model identifier is minimax/minimax-m2.1. It supports tools, so software built to use tool-enabled models can connect it to actions or services.
Its weights are open. People looking for models they can inspect or run themselves can compare it with other open-weight models and other MiniMax models.
Verdict
Choose MiniMax M2.1 if your work centers on programming, written reasoning, or multi-step tasks and you need a large text window. Its open weights distinguish it from models without open weights.
Look elsewhere if you need image understanding or image generation. The model's token billing also separates what you send from what it writes, so costs depend on both sides of each exchange.
Verdict
MiniMax M2.1 fits software projects that need code-focused answers, reasoning, and tool-enabled workflows in text. Its large context and high output limit suit long documents or extended tasks, while open weights distinguish it from models without open weights. It is not a match for image work because both input and output are text. Budgeting is straightforward: input and output have separate rates, with the stated chat example costing about $0.90 for 1,000 turns.
- Open weights are available.
- Supports reasoning and tool use.
- 204,800-token context window.
- Up to 131,072 output tokens.
- Designed for coding and multi-step workflows.
- Text input and output only.
- Input and output are billed at different rates.
- Pick it ifCoding and application-development tasks; Text workflows that use tools; Long prompts needing a large context
- Skip it ifYou need image input or image generation.
Pricing
Input is $0.3 per million tokens and output is $1.2 per million tokens. The rough cost is about 0.019 cents to read a page, about 0.09 cents for one sample chat turn, and about $0.90 for 1,000 such turns.
Questions
What is MiniMax M2.1?
It is a text-only language model from MiniMax for coding, reasoning, tool use, and multi-step workflows.
When was MiniMax M2.1 released?
MiniMax M2.1 was released on December 23, 2025.
Does MiniMax M2.1 have open weights?
Yes. Its weights are open.
How much context can it handle?
It can consider up to 204,800 tokens in one request.
What is the maximum answer length?
Its maximum output is 131,072 tokens.
How much does it cost?
Input costs $0.3 per million tokens and output costs $1.2 per million tokens.
Can it use tools?
Yes. Tool support is listed for the model.
Can it process images?
No. Its listed input and output are text.