TheDrummer_Behemoth-128B-v3-GGUF

Llamacpp imatrix Quantizations of Behemoth-128B-v3 by TheDrummer

Using <a href="https://github.com/ggml-org/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggml-org/llama.cpp/releases/tag/b10472">b10472</a> for quantization.

Original model: https://huggingface.co/TheDrummer/Behemoth-128B-v3

Model details:

• Parameter count: 125B
• Input support: text
• Speculative decoding: no
• imatrix: yes - details
• Perplexity/KLD measured: no

How to run

Prompt format

<s>[SYSTEM_PROMPT]{system_prompt}[/SYSTEM_PROMPT][MODEL_SETTINGS]{"reasoning_effort": "none"}[/MODEL_SETTINGS][INST]{prompt}[/INST]

Don't know which to choose? Grab Q4_K_M (78.41GB) - usually a good mix of size and performance. Download instructions available here

Available files:

Download a specific file:

hf download bartowski/TheDrummer_Behemoth-128B-v3-GGUF --include "TheDrummer_Behemoth-128B-v3-Q4_K_M/*" --local-dir ./

Downloading using the Hugging Face CLI

<details> <summary>Click to view download instructions</summary>

First, make sure you have the Hugging Face CLI installed:

pip install -U "huggingface_hub[cli]"

Download a specific file:

hf download bartowski/TheDrummer_Behemoth-128B-v3-GGUF --include "TheDrummer_Behemoth-128B-v3-Q4_K_M/*" --local-dir ./

The files marked true in the Split column above are stored as multiple parts in a folder. To download all the parts to a local folder, run:

hf download bartowski/TheDrummer_Behemoth-128B-v3-GGUF --include "TheDrummer_Behemoth-128B-v3-Q8_0/*" --local-dir ./

You can either specify a new local-dir (TheDrummer_Behemoth-128B-v3-Q8_0) or download them all in place (./)

</details>

How to run

These quants run with llama.cpp - installable in one line via llama.app:

curl -LsSf https://llama.app/install.sh | sh
llama-server -hf bartowski/TheDrummer_Behemoth-128B-v3-GGUF:Q4_K_M

llama-server includes a built-in chat web UI, served at http://localhost:8080 by default.

These quants were made with llama.cpp release b10472 - if this model's architecture is newly supported, you'll need that release or newer to run them.

They also work in: LM Studio · koboldcpp · ramalama · Jan AI · Text Generation Web UI · LoLLMs · Atomic Chat

imatrix

All quants made using imatrix option, with a calibration corpus rendered through this model's own chat template. The corpus pairs plain prose with tool-calling and reasoning conversations (corpus source data), encoded exactly as this model sees them at inference and processed with --parse-special, so chat-format special tokens contribute to the importance matrix. The corpus rendered for this model is included in this repo: TheDrummer_Behemoth-128B-v3-calibration-v6.txt. The imatrix is available here: TheDrummer_Behemoth-128B-v3-imatrix.gguf.

<details> <summary>Calibration render details</summary>

{
  "generator": "auto_quant_v2 calibration renderer",
  "recipe": "calibration-v6",
  "model": "Behemoth-128B-v3",
  "encoder": "chat_template",
  "chunk_size": 512,
  "prose_chunks": 222,
  "tool_chunks": 395,
  "total_chunks": 617,
  "tool_chunk_fraction": 0.64,
  "n_conversations": 137,
  "extension_convs_used": 0,
  "conversation_token_lengths": [
    721,
    1795,
    1336,
    1633,
    1224,
    1485,
    3267,
    904,
    1333,
    1504,
    1228,
    2270,
    1035,
    1346,
    2902,
    1394,
    1286,
    1146,
    893,
    868,
    1510,
    1139,
    1465,
    1374,
    2006,
    1672,
    1750,
    991,
    1529,
    1762,
    1658,
    1298,
    1396,
    1206,
    1423,
    1812,
    1782,
    1342,
    631,
    2103,
    1591,
    1235,
    1546,
    2141,
    2243,
    1400,
    1794,
    982,
    3057,
    1274,
    3012,
    919,
    1123,
    1073,
    1270,
    854,
    2609,
    989,
    1304,
    1247,
    1308,
    1328,
    1010,
    1320,
    1268,
    1694,
    1040,
    1671,
    2263,
    995,
    682,
    1249,
    3533,
    3006,
    867,
    1014,
    1129,
    1176,
    1396,
    1256,
    1254,
    947,
    1300,
    1162,
    1444,
    1616,
    1543,
    2134,
    999,
    798,
    2854,
    989,
    1509,
    1822,
    2122,
    1371,
    778,
    1505,
    1467,
    1851,
    2013,
    1797,
    988,
    1112,
    1174,
    2925,
    899,
    829,
    914,
    1529,
    1212,
    1715,
    932,
    705,
    669,
    2705,
    1118,
    1261,
    1977,
    2133,
    2796,
    2788,
    913,
    1134,
    996,
    1027,
    1357,
    1126,
    1011,
    1458,
    998,
    867,
    1846,
    1159,
    1082,
    1407,
    1605
  ],
  "warnings": []
}

</details>

Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

ARM/AVX information

llama.cpp automatically "repacks" weights into an interleaved layout at load time for faster inference on ARM and AVX machines - details in this PR. This once required downloading special Q4_0_4_4/4_8/8_8 files; those are long gone. Online repacking now covers Q4_0, IQ4_NL, and most K-quants, so no special quant choice is needed for CPU inference.

Which file should I choose?

<details> <summary>Click here for details</summary>

An older (early 2024) but still useful write-up with charts comparing quant performances is provided by Artefact2 here

The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.

If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.

If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.

Hugging Face can also do this math for you: add your hardware in your Local Apps settings and the model page will show which files fit.

Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.

If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.

If you want to get more into the weeds, you can check out this extremely useful feature chart:

llama.cpp feature matrix

But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.

These I-quants can also be used on CPU, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.

</details>

Credits

Thank you kalomaze and Dampf for assistance in creating the imatrix calibration dataset.

Thank you ZeroWw for the inspiration to experiment with embed/output.

Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski