MI

MiMo V2.6 Flash REAP50

MimoGGUF by patrickbdevaneyUpdated 2026-10-26
Download on Hugging Face
Parameters157B
Context1024K
Layers3
Architecturemimo2
Smallest quant2.4 GB
Quick answer

MiMo V2.6 Flash REAP50 needs about 6.0 GB of memory at BF16 (a 4.5 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 4 quants, from 2.4 GB (Q8_0) to 92.4 GB (MXFP4_MOE). Longer context needs more: at 32K tokens the KV cache alone adds about 4.0 GB.

MiMo V2.6 Flash REAP50 – 157B multimodal MiMo model with 1M token context

MiMo V2.6 Flash REAP50 is a 157 billion‑parameter MiMo model with a 1 048 576‑token context window. It supports vision and audio modalities and includes a 47‑layer MoE architecture with 128 routed experts per layer. The model is released under the MIT license.

Good at

  • Vision and audio multimodal inference
  • Long‑horizon text generation up to 1M tokens
  • Mixed‑modal reasoning with MoE routing

Limits

  • Requires 1 048 576‑token context for full capability
  • No end‑to‑end quality benchmarks published
  • Large VRAM needed for full‑precision quantizations

Best for

  • Multimodal content creation
  • Long‑form document analysis
  • Agentic reasoning tasks

Who it’s for. Developers building multimodal applications that need long context and MoE efficiency. Researchers interested in MiMo architecture and large‑context models. The model is suitable for environments that can provide sufficient VRAM and support llama.cpp multimodal pipelines.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q8_02.4 GB 3.9 GB
Fits Get file →
BF164.5 GB 6.0 GB
Fits Get file →
Q2_K66.2 GB 67.7 GB
Too big Get file →
MXFP4_MOE92.4 GB 93.9 GB
Too big Get file →

For 8 GB VRAM use Q8_0 (2.38 GB, ~3.9 GB memory at 8 k context). For 16 GB VRAM use BF16 (4.48 GB, ~6 GB memory). Larger budgets can use Q2_K (66.19 GB) or MXFP4_MOE (92.4 GB) for higher fidelity.

What GPU do you need for MiMo V2.6 Flash REAP50?

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: the best quant that fits is Q8_0 (2.4 GB file, about 3.9 GB in use).
  • 6 GB VRAM: the best quant that fits is BF16 (4.5 GB file, about 6.0 GB in use).
  • 8 GB VRAM: the best quant that fits is BF16 (4.5 GB file, about 6.0 GB in use).
  • 12 GB VRAM: the best quant that fits is BF16 (4.5 GB file, about 6.0 GB in use).
  • 16 GB VRAM: the best quant that fits is BF16 (4.5 GB file, about 6.0 GB in use).
  • 24 GB VRAM: the best quant that fits is BF16 (4.5 GB file, about 6.0 GB in use).
  • 32 GB VRAM: the best quant that fits is BF16 (4.5 GB file, about 6.0 GB in use).
  • 48 GB VRAM: the best quant that fits is BF16 (4.5 GB file, about 6.0 GB in use).

Which MiMo V2.6 Flash REAP50 quant should you choose?

BF16 (4.5 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 Q8_0 at 2.4 GB, with the quality trade-offs that come with fewer bits. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturemimo2
Parameters157B (approximate, mixture of experts)
Layers3
Trained context1,048,576 tokens
KV cache (fp16)120 KB per token

Source

Quantized bypatrickbdevaney
FamilyMimo
Licensemit
Downloads (HF)3,045
Last updated2026-10-26

Frequently asked questions about MiMo V2.6 Flash REAP50

How much VRAM does MiMo V2.6 Flash REAP50 need?

About 6.0 GB at BF16 with an 8K context window. That is the 4.5 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 MiMo V2.6 Flash REAP50 on an 8 GB GPU?

Yes, with BF16: the 4.5 GB file needs about 6.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of MiMo V2.6 Flash REAP50 GGUF?

It depends on the quant: Q8_0 is 2.4 GB, BF16 is 4.5 GB, Q2_K is 66.2 GB, MXFP4_MOE is 92.4 GB.

What context length does MiMo V2.6 Flash REAP50 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 1.0 GB at 8K and 4.0 GB at 32K.

Where can I download MiMo V2.6 Flash REAP50 GGUF?

The files are on Hugging Face in the patrickbdevaney/MiMo-V2.6-Flash-REAP50-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 MiMo V2.6 Flash REAP50 use?

Hugging Face lists the license as mit. 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 MiMo V2.6 Flash REAP50 on 8 GB VRAM?

Yes, the Q8_0 quantization is 2.38 GB and uses about 3.9 GB of VRAM at 8 k context.

Does the model support vision?

Yes, it includes a 28‑layer ViT vision encoder and multimodal projectors for vision and audio.

Is there a benchmark score available?

No end‑to‑end quality benchmark has been published for these GGUF files.

Other GGUFs of MiMo-V2.6-Flash-RL

Similar models

Browse all →

More from patrickbdevaney

All by patrickbdevaney →

More Mimo models

All Mimo →

Other large models

All large models →