# MiniMax M2.1

> Open-weight MoE coding model built for multi-language programming and real-world agentic tasks.

- Maker: MiniMax
- Line: MiniMax M-Series
- Canonical page: https://ai-tldr.dev/models/minimax-m2-1/
- Updated: 2026-06-18

## In short

MiniMax M2.1 is an open-weight large language model from MiniMax, released on December 23, 2025 as an enhanced revision of MiniMax M2. It has a 205K-token context window and costs $0.30 per million input tokens and $1.20 per million output tokens. You can use it through MiniMax and OpenRouter.

## Facts

| Fact | Value | Source |
| --- | --- | --- |
| Context window | 205K (max output 196,608 tokens) | [1] |
| Input price | $0.30 / 1M tokens | [2] |
| Output price | $1.20 / 1M tokens | [2] |
| Licence | Modified MIT · open weights | [1] |
| Released | 23 December 2025 | [3] |
| Parameters | 230B total / 10B active (MoE) | [1] |

## Sources

1. [Hugging Face model card (MiniMaxAI/MiniMax-M2.1)](https://huggingface.co/MiniMaxAI/MiniMax-M2.1)
2. [MiniMax pay-as-you-go pricing](https://platform.minimax.io/docs/guides/pricing-paygo)
3. [Official announcement: MiniMax M2.1](https://www.minimax.io/news/minimax-m21)

## Overview

MiniMax M2.1 is an open-weight large language model from MiniMax, released on December 23, 2025 as an enhanced revision of MiniMax M2. It is a sparse Mixture-of-Experts (MoE) model with roughly 230 billion total parameters but only about 10 billion activated per token, which keeps inference fast and cheap while retaining the capacity of a much larger dense model. The weights ship on Hugging Face under a modified MIT license, so teams can run M2.1 locally with SGLang, vLLM, Transformers, or KTransformers instead of relying solely on the hosted API.

The headline focus of MiniMax M2.1 is multi-language programming. MiniMax says it systematically improved performance across Rust, Java, Golang, C++, Kotlin, Objective-C, TypeScript, and JavaScript, spanning low-level systems code through web and mobile application development. It is positioned as a 'digital employee' for end-to-end coding and agentic workflows, building on MiniMax's interleaved-thinking approach for multi-step problem solving, precise refactoring, and tool use.

On benchmarks, MiniMax M2.1 reaches 74% on SWE-bench Verified and 72.5% on SWE-bench Multilingual, which MiniMax reports as competitive with or ahead of Claude Sonnet 4.5 on coding-heavy tasks, at a fraction of the cost. With a roughly 205K-token context window, text-only inputs, and hosted pricing of $0.30 per million input tokens and $1.20 per million output tokens, M2.1 targets developers who want strong agentic coding without frontier-model API bills.

## Benchmarks

- SWE-bench Verified: 74% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- SWE-bench Multilingual: 72.5% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- Multi-SWE-bench: 49.4% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- Terminal-Bench 2.0: 47.9% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- VIBE (average): 88.6% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- Toolathlon: 43.5% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- BrowseComp: 47.4% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- AIME 2025: 83% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- MMLU-Pro: 88% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))
- GPQA Diamond: 83% ([source](https://huggingface.co/MiniMaxAI/MiniMax-M2.1))

## Pricing

- Input: $0.30 / 1M tokens per 1M tokens
- Output: $1.20 / 1M tokens per 1M tokens

Source: https://platform.minimax.io/docs/guides/pricing-paygo

## Strengths

- Strong real-world software-engineering performance: 74% on SWE-bench Verified, competitive with leading proprietary coding models
- Broad multi-language programming coverage tuned for Rust, Java, Golang, C++, Kotlin, Objective-C, TypeScript, and JavaScript
- Open weights under a modified MIT license, enabling local and self-hosted deployment via SGLang, vLLM, Transformers, and KTransformers
- Efficient sparse MoE design (only ~10B of ~230B parameters active per token) keeps inference fast and inexpensive
- Low hosted pricing at $0.30 / $1.20 per million input/output tokens, far below frontier proprietary coding models
- Built for agentic, multi-step tasks with interleaved thinking and tool use

## Best for

- Autonomous and assisted software engineering across many programming languages
- Agentic coding workflows that plan, edit, run tools, and iterate over multi-step tasks
- Code refactoring and migration in large, multi-language codebases
- Web and mobile (Android/iOS) application development assistance
- Self-hosted or on-premise coding assistants where open weights and data control matter
- Cost-sensitive automation: documentation, data handling, and multi-step pipelines at low per-token prices

## How to access

- [MiniMax](https://platform.minimax.io/docs/release-notes/models) — model id `MiniMax-M2.1`
- [OpenRouter](https://openrouter.ai/minimax/minimax-m2.1) — model id `minimax/minimax-m2.1`

## MiniMax M-Series versions

- [MiniMax M3](https://ai-tldr.dev/models/minimax-m3/) (1 June 2026)
- [MiniMax M2.7 / M2.7-highspeed](https://ai-tldr.dev/models/minimax-m2-7/) (18 March 2026)
- [MiniMax M2.5 / M2.5-Lightning](https://ai-tldr.dev/models/minimax-m2-5/) (12 February 2026)
- MiniMax M2.1 (23 December 2025) — this page
- [MiniMax M2](https://ai-tldr.dev/models/minimax-m2/) (27 October 2025)

## FAQ

### Is MiniMax M2.1 open source?

The model weights are open and published on Hugging Face under a modified MIT license, so you can download and run MiniMax M2.1 locally with frameworks like SGLang, vLLM, Transformers, or KTransformers, in addition to using MiniMax's hosted API.

### How much does MiniMax M2.1 cost?

On MiniMax's hosted platform, M2.1 is priced at $0.30 per million input tokens and $1.20 per million output tokens, which is substantially cheaper than most frontier proprietary coding models.

### How good is MiniMax M2.1 at coding?

MiniMax M2.1 scores 74% on SWE-bench Verified and 72.5% on SWE-bench Multilingual, placing it among the stronger coding models. MiniMax positions it as competitive with Claude Sonnet 4.5 on software-engineering tasks while costing far less.

### How big is MiniMax M2.1 and what is its context window?

M2.1 is a Mixture-of-Experts model with roughly 230 billion total parameters but only about 10 billion active per token (256 experts, 8 activated). It supports a context window of about 205K tokens with up to 196,608 output tokens, and it processes text inputs.

## Links

- [Official announcement: MiniMax M2.1](https://www.minimax.io/news/minimax-m21)
- [Hugging Face model card (MiniMaxAI/MiniMax-M2.1)](https://huggingface.co/MiniMaxAI/MiniMax-M2.1)
- [MiniMax API model release notes](https://platform.minimax.io/docs/release-notes/models)
- [MiniMax pay-as-you-go pricing](https://platform.minimax.io/docs/guides/pricing-paygo)
- [OpenRouter: MiniMax M2.1](https://openrouter.ai/minimax/minimax-m2.1)
- [Artificial Analysis: MiniMax M2.1 providers](https://artificialanalysis.ai/models/minimax-m2-1/providers)

Source: AI/TLDR LLM registry, https://ai-tldr.dev/models/minimax-m2-1/