MI

MiniMax M3.1 Flash Q2 K

MinimaxGGUF by smalininUpdated 2026-10-24
Download on Hugging Face
Parameters2.12B
Context1024K
Layers5
Architecturedflash
Smallest quant3.9 GB
Quick answer

MiniMax M3.1 Flash Q2 K needs about 4.8 GB of memory at BF16 (a 4.3 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 3 quants, from 3.9 GB (BF16) to 162.3 GB (Q2_K). Longer context needs more: at 32K tokens the KV cache alone adds about 0.3 GB.

MiniMax M3.1 Flash Q2_K – a mixed‑precision GGUF model with vision, reasoning, and tools support

MiniMax M3.1 Flash Q2_K is a 2.12 billion‑parameter model quantized with a custom Q2_K‑Protected recipe. It supports text, image, and video inputs and includes a separate vision projector and DSpark speculative decoding model. The model uses MiniMax Sparse Attention and is available in three GGUF files.

Good at

  • Vision and image understanding
  • Reasoning tasks
  • Tool usage

Limits

  • Requires large VRAM – 162.8 GB at 8 k context for Q2_K
  • No explicit license details
  • No separate language list provided

Best for

  • Multimodal content generation
  • Complex reasoning workflows
  • Tool‑integrated applications

Who it’s for. Developers and researchers needing a multimodal LLM with vision, reasoning, and tool capabilities. The model is suitable for projects that can accommodate its high memory requirements and custom quantization.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
BF163.9 GB 4.5 GB
Fits Get file →
BF164.3 GB 4.8 GB
Fits Get file →
Q2_K162.3 GB (6 parts) 162.8 GB
Too big Get file →

For 8 k context, the Q2_K quantization needs about 162.8 GB of VRAM. If limited to 4 GB VRAM, use the BF16 files: the 3.93 GB file requires roughly 4.5 GB of VRAM, and the 4.26 GB file requires about 4.8 GB. These are the only quants available in the file list.

What GPU do you need for MiniMax M3.1 Flash Q2 K?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: the best quant that fits is BF16 (4.3 GB file, about 4.8 GB in use).
  • 8 GB VRAM: the best quant that fits is BF16 (4.3 GB file, about 4.8 GB in use).
  • 12 GB VRAM: the best quant that fits is BF16 (4.3 GB file, about 4.8 GB in use).
  • 16 GB VRAM: the best quant that fits is BF16 (4.3 GB file, about 4.8 GB in use).
  • 24 GB VRAM: the best quant that fits is BF16 (4.3 GB file, about 4.8 GB in use).
  • 32 GB VRAM: the best quant that fits is BF16 (4.3 GB file, about 4.8 GB in use).
  • 48 GB VRAM: the best quant that fits is BF16 (4.3 GB file, about 4.8 GB in use).

Which MiniMax M3.1 Flash Q2 K quant should you choose?

BF16 (4.3 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.

If memory is tight, the smallest file is BF16 at 3.9 GB, with the quality trade-offs that come with fewer bits. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturedflash
Parameters2.12B
Layers5
Trained context1,048,576 tokens
KV cache (fp16)10 KB per token

Source

Quantized bysmalinin
FamilyMinimax
Licenseother
Downloads (HF)1,090
Last updated2026-10-24

Frequently asked questions about MiniMax M3.1 Flash Q2 K

How much VRAM does MiniMax M3.1 Flash Q2 K need?

About 4.8 GB at BF16 with an 8K context window. That is the 4.3 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run MiniMax M3.1 Flash Q2 K on an 8 GB GPU?

Yes, with BF16: the 4.3 GB file needs about 4.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of MiniMax M3.1 Flash Q2 K GGUF?

It depends on the quant: BF16 is 3.9 GB, BF16 is 4.3 GB, Q2_K is 162.3 GB.

What context length does MiniMax M3.1 Flash Q2 K support?

The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 0.1 GB at 8K and 0.3 GB at 32K.

Where can I download MiniMax M3.1 Flash Q2 K GGUF?

The files are on Hugging Face in the smalinin/MiniMax-M3.1-Flash-Q2_K-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does MiniMax M3.1 Flash Q2 K use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run this model on 8 GB VRAM?

No. The Q2_K quantization requires about 162.8 GB of VRAM at 8 k context. The BF16 variants need 4.5–4.8 GB, which is still above 8 GB.

Does the model support text only?

No. The model accepts text, images, and video inputs and includes a vision projector.

Is there a license restriction?

The license is listed as "other"; consult the original repository for details.

Similar models

Browse all →

More from smalinin

All by smalinin →

More Minimax models

All Minimax →

Other tiny models

All tiny models →