> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/automodel/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo/automodel/_mcp/server.

# MiniMax-M3

[MiniMaxAI/MiniMax-M3](https://huggingface.co/MiniMaxAI/MiniMax-M3) is MiniMaxAI's 428B A22B Mixture-of-Experts first vision-language model combining long-context reasoning, agentic workflows, and creative capabilities in a single platform.

:::\{card}

|                     |                                                                               |
| ------------------- | ----------------------------------------------------------------------------- |
| **Task**            | Image-Text-to-Text / Video-Text-to-Text                                       |
| **Architecture**    | `MiniMaxM3SparseForConditionalGeneration` — 428B total / 22B active MoE VLM   |
| **Language Module** | MiniMax-M2.7 backbone, 60 layers, 128 experts, MiniMax Sparse Attention (MSA) |
| **Vision Module**   | CLIP-style ViT, 32 layers, image size 336×336 up to 2016×2016                 |
| **Context Window**  | 512k tokens                                                                   |
| **Precision**       | BF16 and MXFP8                                                                |
| **HF Org**          | [MiniMaxAI](https://huggingface.co/MiniMaxAI)                                 |
| :::                 |                                                                               |

## Positioning

MiniMax-M3 targets advanced use cases such as long-form video understanding, extended coding tasks (8+ hours), and high-quality design workflows supporting upto 1M context length.

## Architecture

* **Language backbone:** derived from  MiniMax-M2.7 with 60 layers (3 dense + 57 MoE), 64 attention heads, 128 experts, block-sparse attention called MiniMax Sparse Attention (MSA) on the MoE layers and a 512k context length.
* **Vision backbone:** CLIP-style ViT with 32 layers and dynamic resolution image input from 336×336 up to 2016×2016.
* **Optimization target:** trained on Hopper GPUs, with BF16 and FP8 support.

## Key Strengths

* **Ultra-long context via MSA (MiniMax Sparse Attention).** A block-sparse attention design that avoids full attention’s quadratic cost enabling up to 1M-token context.
* **Frontier coding and agentic performance.** Positioned as a coding and agent-first model, with strong benchmark numbers and good at multi-turn, collaborative agent behavior.
* **Native multimodality (image, video, desktop use).** It supports image and video input and can operate a desktop computer.
* **Long-horizon autonomous execution on complex real-world tasks.** Can work over hours to a full day without human intervention.

## Available Models

* **MiniMax-M3** — registered as `MiniMaxM3SparseForConditionalGeneration` model class.

## Example HF Models

| Model      | HF ID                                                                 |
| ---------- | --------------------------------------------------------------------- |
| MiniMax-M3 | [`MiniMaxAI/MiniMax-M3`](https://huggingface.co/MiniMaxAI/MiniMax-M3) |

## Example Recipes

* [Full SFT — MedPix, EP32 + PP4](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/minimax_m3/minimax_m3_vl_sft_ep32pp4.yaml)
* [Full SFT — MedPix CP2 comparison](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/minimax_m3/minimax_m3_vl_sft_cp2_medpix_2k.yaml)
* [Full SFT — Tulu 3 text, CP8](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/minimax_m3/minimax_m3_vl_sft_tulu3_text_cp8_16k.yaml)
* [LoRA — MedPix, PP4 + EP8](https://github.com/NVIDIA-NeMo/Automodel/blob/main/examples/vlm_finetune/minimax_m3/minimax_m3_vl_lora_pp4ep8_8node.yaml)

See the [MiniMax-M3 fine-tuning guide](/recipes-e2e-examples/minimax-m3) for the training setup and launch notes.

## Hugging Face Model Cards

* [MiniMaxAI/MiniMax-M3](https://huggingface.co/MiniMaxAI/MiniMax-M3)