Mellum2.1 12B A2.5B MaxQuality iMatrix needs about 9.0 GB of memory at Q4_K_M (a 8.1 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 2 quants, from 6.3 GB (Q3_K_M) to 8.1 GB (Q4_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 1.9 GB.
Mellum2.1 12B A2.5B MaxQuality iMatrix – a 12B MoE model for agentic coding, math and reasoning
Mellum2.1 12B A2.5B MaxQuality iMatrix is a 12B total, 2.5B active parameter mixture‑of‑experts model. It supports a context length of 131,072 tokens and is trained for agentic tasks such as repository navigation, tool use, and complex reasoning.
Good at
- Agentic coding and tool use
- Complex reasoning and math
- Large context handling
Limits
- Requires high VRAM for MoE execution
- No vision support
- Licensed under Apache 2.0
Best for
- Software engineering agents
- Math problem solving
- Long‑form reasoning tasks
Who it’s for. Developers and researchers needing a large‑context, agentic LLM for coding, math, and reasoning. The model is suitable for environments that can accommodate its VRAM and CPU requirements.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_M | 6.3 GB | 7.3 GB | Tight | Get file → | |
| Q4_K_M | 8.1 GB | 9.0 GB | Too big | Get file → |
For 16 GB VRAM, use Q4_K_M (8.07 GB). For 8 GB VRAM, use Q3_K_M (6.33 GB). Memory estimates are approximate and based on 8K context usage.
What GPU do you need for Mellum2.1 12B A2.5B MaxQuality iMatrix?
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: the best quant that fits is Q3_K_M (6.3 GB file, about 7.3 GB in use).
- 12 GB VRAM: the best quant that fits is Q4_K_M (8.1 GB file, about 9.0 GB in use).
Which Mellum2.1 12B A2.5B MaxQuality iMatrix quant should you choose?
Q4_K_M is the usual starting point for Mellum2.1 12B A2.5B MaxQuality iMatrix: it balances file size and quality well. Here it is a 8.1 GB file.
If memory is tight, the smallest file is Q3_K_M at 6.3 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Mellum2.1 12B A2.5B MaxQuality iMatrix
How much VRAM does Mellum2.1 12B A2.5B MaxQuality iMatrix need?
About 9.0 GB at Q4_K_M with an 8K context window. That is the 8.1 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 Mellum2.1 12B A2.5B MaxQuality iMatrix on an 8 GB GPU?
Yes, with Q3_K_M: the 6.3 GB file needs about 7.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Mellum2.1 12B A2.5B MaxQuality iMatrix GGUF?
It depends on the quant: Q3_K_M is 6.3 GB, Q4_K_M is 8.1 GB.
What context length does Mellum2.1 12B A2.5B MaxQuality iMatrix support?
The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.5 GB at 8K and 1.9 GB at 32K.
Where can I download Mellum2.1 12B A2.5B MaxQuality iMatrix GGUF?
The files are on Hugging Face in the offmonreal/Mellum2.1-12B-A2.5B-MaxQuality-iMatrix-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 Mellum2.1 12B A2.5B MaxQuality iMatrix use?
Hugging Face lists the license as apache-2.0. 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 Mellum2.1 on a 8 GB VRAM GPU?
Yes, by using the Q3_K_M quant which is 6.33 GB, leaving room for KV cache and overhead.
Does this model support vision tasks?
No, the model has no vision capabilities.
What license applies to this model?
The model is released under the Apache 2.0 license.