MiniMaxAI / MiniMax-M3 like 795 Follow MiniMax 6.9k

---

MiniMax-M3 is a native multimodal model with 1M context. It has \~428B parameters and \~23B activated parameters.

Highlights:

• Native Multimodality: M3 undergoes mixed-modality training from the very first step, enabling deeper semantic fusion across text, image, and video.
• Context Scaling via Sparse Attention: M3 introduces MiniMax Sparse Attention (MSA) to improve long context efficiency. M3 delivers 9× prefill and 15× decode speedups compared to M2 at 1M context, reducing per-token compute to 1/20.
• Coding & Cowork Capability: M3 achieves frontier-level performance across long-horizon agentic benchmarks, excelling in both coding and cowork.



MiniMax Sparse Attention (MSA)

M3 is powered by MiniMax Sparse Attention (MSA), a high-performance sparse attention operator designed for million-token contexts. Compared with GQA, MSA dramatically reduces the attention compute and memory footprint while preserving model quality.



📄 Read the technical report: arXiv:2606.13392 · Hugging Face Papers

How to Use

• MiniMax Agent
• MiniMax API

M3 supports three reasoning modes through the `thinking` parameter:

• `enabled` — Reasoning is always enabled.
• `adaptive` — M3 automatically determines when additional reasoning is beneficial.
• `disabled` — Reasoning is disabled to minimize latency and maximize throughput.

Local Deployment

Download the model:

hf download MiniMaxAI/MiniMax-M3 --local-dir MiniMax-M3

We recommend the following inference frameworks (listed alphabetically) to serve the model:

• SGLang \- see SGLang cookbook.
• vLLM \- see vLLM recipes.
• Transformers \- see Transformers docs.

Inference Parameters

We recommend the following parameters for best performance: `temperature=1.0`, `top_p=0.95`, `top_k=40`.

Contact Us

Contact us at model@minimax.io.

Downloads last month 14,312

Safetensors[](https://huggingface.co/docs/safetensors)

Model size

427B params

Tensor type

BF16

·

F32

·

Chat template

Files info

Inference Providers NEW

Novita

• +1

Image-Text-to-Text

Examples

Input a message to start chatting with MiniMaxAI/MiniMax-M3.

Send

View Code Snippets

Compare providers

Model tree for MiniMaxAI/MiniMax-M3 [](https://huggingface.co/docs/hub/model-cards#specifying-a-base-model)

Finetunes

1 model

Quantizations

[](https://huggingface.co/models?apps=llama.cpp&other=base%5Fmodel:quantized:MiniMaxAI/MiniMax-M3 "Use with llama.cpp")[](https://huggingface.co/models?apps=lmstudio&other=base%5Fmodel:quantized:MiniMaxAI/MiniMax-M3 "Use with LM Studio")[](https://huggingface.co/models?apps=jan&other=base%5Fmodel:quantized:MiniMaxAI/MiniMax-M3 "Use with Jan")[](https://huggingface.co/models?apps=ollama&other=base%5Fmodel:quantized:MiniMaxAI/MiniMax-M3 "Use with Ollama")

18 models

Spaces using MiniMaxAI/MiniMax-M3 7

Collection including MiniMaxAI/MiniMax-M3

3 items • Updated about 16 hours ago • 4

Paper for MiniMaxAI/MiniMax-M3

Paper • 2606.13392 • Published 5 days ago • 130