MiMo-V2.6-Flash-RL needs about 22.1 GB of memory at BF16 (a 4.5 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 7 quants, from 1.3 GB (Q4_0) to 167.4 GB (MXFP4). Longer context needs more: at 32K tokens the KV cache alone adds about 68.5 GB.
MiMo‑V2.6‑Flash‑RL: a large multimodal RL‑trained model
MiMo‑V2.6‑Flash‑RL is a 309 billion‑parameter model trained with reinforcement learning. It supports text, image, video, and audio and can handle up to 1,048,576 context tokens. The model is released under the MIT license.
Good at
- Multimodal inference
- Long‑horizon reasoning
- Reinforcement‑learning‑derived behavior
Limits
- Requires large VRAM for full precision
- No separate vision license constraints
- Limited to 1M context tokens
Best for
- Agent‑based task automation
- Multimodal content generation
- Long‑form dialogue
Who it’s for. Developers building multimodal agents or long‑context applications. Researchers needing a large RL‑trained baseline. Users with GPUs that can accommodate the memory footprint.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_0 | 1.3 GB | 18.9 GB | Too big | Get file → | |
| Q8_0 | 2.4 GB | 20.0 GB | Too big | Get file → | |
| Q8_0 | 2.9 GB | 20.5 GB | Too big | Get file → | |
| BF16 | 4.5 GB | 22.1 GB | Too big | Get file → | |
| BF16 | 5.4 GB | 23.1 GB | Too big | Get file → | |
| Q2_K | 126.4 GB (2 parts) | 144.0 GB | Too big | Get file → | |
| MXFP4 | 167.4 GB (2 parts) | 185.0 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_0 (1.26 GB file, 18.9 GB memory at 8 k context). For 16 GB VRAM, Q8_0 (2.38 GB file, 20 GB memory) is suitable. For 32 GB VRAM, BF16 (4.48 GB file, 22.1 GB memory) offers higher precision.
What GPU do you need for MiMo-V2.6-Flash-RL?
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: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: the best quant that fits is BF16 (5.4 GB file, about 23.1 GB in use).
- 32 GB VRAM: the best quant that fits is BF16 (5.4 GB file, about 23.1 GB in use).
- 48 GB VRAM: the best quant that fits is BF16 (5.4 GB file, about 23.1 GB in use).
Which MiMo-V2.6-Flash-RL 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 Q4_0 at 1.3 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about MiMo-V2.6-Flash-RL
How much VRAM does MiMo-V2.6-Flash-RL need?
About 22.1 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-RL on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 18.9 GB, so you would need a larger GPU or CPU offloading.
What is the file size of MiMo-V2.6-Flash-RL GGUF?
It depends on the quant: Q4_0 is 1.3 GB, Q8_0 is 2.4 GB, Q8_0 is 2.9 GB, BF16 is 4.5 GB, BF16 is 5.4 GB, Q2_K is 126.4 GB and 1 more.
What context length does MiMo-V2.6-Flash-RL 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 17.1 GB at 8K and 68.5 GB at 32K.
Where can I download MiMo-V2.6-Flash-RL GGUF?
The files are on Hugging Face in the ggml-org/MiMo-V2.6-Flash-RL-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-RL 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‑RL on 8 GB VRAM?
Yes, using the Q4_0 quantization which requires about 18.9 GB memory at 8 k context, so it fits on 8 GB VRAM with careful batching.
Does the model support vision tasks?
Yes, it includes vision and audio encoders and is detected to have vision ability.
What license applies to this model?
The model is released under the MIT license.