GRM 2.6 Plus 0628 needs about 16.8 GB of memory at IQ4_XS (a 14.2 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 16 quants, from 8.9 GB (IQ2_S) to 24.1 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.
GRM 2.6 Plus 0628 is a 27.3B-parameter Grm model in GGUF format, quantized by NikiKrutan. Its trained context length is 256K tokens. The repository has 16 quant files, from 8.9 GB (IQ2_S) to 24.1 GB (Q6_K).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ2_S | 8.9 GB | 11.6 GB | Too big | Get file → | |
| IQ2_S | 9.5 GB | 12.2 GB | Too big | Get file → | |
| IQ3_XXS | 10.2 GB | 12.9 GB | Too big | Get file → | |
| IQ3_XXS | 10.9 GB | 13.6 GB | Too big | Get file → | |
| IQ3_S | 11.6 GB | 14.3 GB | Too big | Get file → | |
| IQ3_S | 12.4 GB | 15.1 GB | Too big | Get file → | |
| IQ4_XS | 13.3 GB | 16.0 GB | Too big | Get file → | |
| IQ4_XS | 14.2 GB | 16.8 GB | Too big | Get file → | |
| IQ4_XS | 15.2 GB | 17.9 GB | Too big | Get file → | |
| IQ4_XS | 16.1 GB | 18.8 GB | Too big | Get file → | |
| Q5_K_M | 17.3 GB | 20.0 GB | Too big | Get file → | |
| Q5_K_M | 18.5 GB | 21.1 GB | Too big | Get file → | |
| Q5_K_M | 19.7 GB | 22.4 GB | Too big | Get file → | |
| Q6_K | 21.1 GB | 23.8 GB | Too big | Get file → | |
| Q6_K | 22.5 GB | 25.2 GB | Too big | Get file → | |
| Q6_K | 24.1 GB | 26.8 GB | Too big | Get file → |
What GPU do you need for GRM 2.6 Plus 0628?
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: the best quant that fits is IQ2_S (8.9 GB file, about 11.6 GB in use).
- 16 GB VRAM: the best quant that fits is IQ4_XS (13.3 GB file, about 16.0 GB in use).
- 24 GB VRAM: the best quant that fits is Q6_K (21.1 GB file, about 23.8 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K (24.1 GB file, about 26.8 GB in use).
Which GRM 2.6 Plus 0628 quant should you choose?
IQ4_XS (14.2 GB) is the middle option in this repository. The usual sweet spot between file size and quality, and the quant most people start with.
If memory is tight, the smallest file is IQ2_S at 8.9 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about GRM 2.6 Plus 0628
How much VRAM does GRM 2.6 Plus 0628 need?
About 16.8 GB at IQ4_XS with an 8K context window. That is the 14.2 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 GRM 2.6 Plus 0628 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 11.6 GB, so you would need a larger GPU or CPU offloading.
What is the file size of GRM 2.6 Plus 0628 GGUF?
It depends on the quant: IQ2_S is 8.9 GB, IQ2_S is 9.5 GB, IQ3_XXS is 10.2 GB, IQ3_XXS is 10.9 GB, IQ3_S is 11.6 GB, IQ3_S is 12.4 GB and 10 more.
What context length does GRM 2.6 Plus 0628 support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.2 GB at 8K and 8.7 GB at 32K.
Where can I download GRM 2.6 Plus 0628 GGUF?
The files are on Hugging Face in the NikiKrutan/GRM-2.6-Plus-0628-MTP-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 GRM 2.6 Plus 0628 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.