bartowski
/

Qwen_Qwen3-30B-A3B-GGUF

@@ -1,11 +1,175 @@
 ---
 quantized_by: bartowski
 pipeline_tag: text-generation
-license: apache-2.0
-base_model: Qwen/Qwen3-30B-A3B
-base_model_relation: quantized
 ---
-llama.cpp quants of https://huggingface.co/Qwen/Qwen3-30B-A3B
-Fixed versions are currently going up! All existing files have been replaced with proper ones, enjoy :)

 ---
 quantized_by: bartowski
 pipeline_tag: text-generation
 ---
+## Llamacpp imatrix Quantizations of Qwen3-30B-A3B by Qwen
+Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b5200">b5200</a> for quantization.
+Original model: https://huggingface.co/Qwen/Qwen3-30B-A3B
+All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8)
+Run them in [LM Studio](https://lmstudio.ai/)
+Run them directly with [llama.cpp](https://github.com/ggerganov/llama.cpp), or any other llama.cpp based project
+## Prompt format
+```
+<|im_start|>system
+{system_prompt}<|im_end|>
+<|im_start|>user
+{prompt}<|im_end|>
+<|im_start|>assistant
+```
+## What's new:
+Improved imatrix dataset
+## Download a file (not the whole branch) from below:
+| Filename | Quant type | File Size | Split | Description |
+| -------- | ---------- | --------- | ----- | ----------- |
+| [Qwen3-30B-A3B-bf16.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/tree/main/Qwen_Qwen3-30B-A3B-bf16) | bf16 | 61.10GB | true | Full BF16 weights. |
+| [Qwen3-30B-A3B-Q8_0.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q8_0.gguf) | Q8_0 | 32.48GB | false | Extremely high quality, generally unneeded but max available quant. |
+| [Qwen3-30B-A3B-Q6_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q6_K_L.gguf) | Q6_K_L | 25.26GB | false | Uses Q8_0 for embed and output weights. Very high quality, near perfect, *recommended*. |
+| [Qwen3-30B-A3B-Q6_K.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q6_K.gguf) | Q6_K | 25.10GB | false | Very high quality, near perfect, *recommended*. |
+| [Qwen3-30B-A3B-Q5_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q5_K_L.gguf) | Q5_K_L | 21.94GB | false | Uses Q8_0 for embed and output weights. High quality, *recommended*. |
+| [Qwen3-30B-A3B-Q5_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q5_K_M.gguf) | Q5_K_M | 21.74GB | false | High quality, *recommended*. |
+| [Qwen3-30B-A3B-Q5_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q5_K_S.gguf) | Q5_K_S | 21.10GB | false | High quality, *recommended*. |
+| [Qwen3-30B-A3B-Q4_1.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q4_1.gguf) | Q4_1 | 19.21GB | false | Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon. |
+| [Qwen3-30B-A3B-Q4_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q4_K_L.gguf) | Q4_K_L | 18.86GB | false | Uses Q8_0 for embed and output weights. Good quality, *recommended*. |
+| [Qwen3-30B-A3B-Q4_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q4_K_M.gguf) | Q4_K_M | 18.63GB | false | Good quality, default size for most use cases, *recommended*. |
+| [Qwen3-30B-A3B-Q4_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q4_K_S.gguf) | Q4_K_S | 17.98GB | false | Slightly lower quality with more space savings, *recommended*. |
+| [Qwen3-30B-A3B-Q4_0.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q4_0.gguf) | Q4_0 | 17.63GB | false | Legacy format, offers online repacking for ARM and AVX CPU inference. |
+| [Qwen3-30B-A3B-IQ4_NL.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ4_NL.gguf) | IQ4_NL | 17.39GB | false | Similar to IQ4_XS, but slightly larger. Offers online repacking for ARM CPU inference. |
+| [Qwen3-30B-A3B-IQ4_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ4_XS.gguf) | IQ4_XS | 16.46GB | false | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. |
+| [Qwen3-30B-A3B-Q3_K_XL.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q3_K_XL.gguf) | Q3_K_XL | 14.86GB | false | Uses Q8_0 for embed and output weights. Lower quality but usable, good for low RAM availability. |
+| [Qwen3-30B-A3B-Q3_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q3_K_L.gguf) | Q3_K_L | 14.58GB | false | Lower quality but usable, good for low RAM availability. |
+| [Qwen3-30B-A3B-Q3_K_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q3_K_M.gguf) | Q3_K_M | 14.08GB | false | Low quality. |
+| [Qwen3-30B-A3B-IQ3_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ3_M.gguf) | IQ3_M | 14.08GB | false | Medium-low quality, new method with decent performance comparable to Q3_K_M. |
+| [Qwen3-30B-A3B-Q3_K_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q3_K_S.gguf) | Q3_K_S | 13.43GB | false | Low quality, not recommended. |
+| [Qwen3-30B-A3B-IQ3_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ3_XS.gguf) | IQ3_XS | 12.74GB | false | Lower quality, new method with decent performance, slightly better than Q3_K_S. |
+| [Qwen3-30B-A3B-IQ3_XXS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ3_XXS.gguf) | IQ3_XXS | 12.22GB | false | Lower quality, new method with decent performance, comparable to Q3 quants. |
+| [Qwen3-30B-A3B-Q2_K_L.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q2_K_L.gguf) | Q2_K_L | 11.21GB | false | Uses Q8_0 for embed and output weights. Very low quality but surprisingly usable. |
+| [Qwen3-30B-A3B-Q2_K.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-Q2_K.gguf) | Q2_K | 10.91GB | false | Very low quality but surprisingly usable. |
+| [Qwen3-30B-A3B-IQ2_M.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ2_M.gguf) | IQ2_M | 10.43GB | false | Relatively low quality, uses SOTA techniques to be surprisingly usable. |
+| [Qwen3-30B-A3B-IQ2_S.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ2_S.gguf) | IQ2_S | 9.22GB | false | Low quality, uses SOTA techniques to be usable. |
+| [Qwen3-30B-A3B-IQ2_XS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ2_XS.gguf) | IQ2_XS | 9.14GB | false | Low quality, uses SOTA techniques to be usable. |
+| [Qwen3-30B-A3B-IQ2_XXS.gguf](https://huggingface.co/bartowski/Qwen_Qwen3-30B-A3B-GGUF/blob/main/Qwen_Qwen3-30B-A3B-IQ2_XXS.gguf) | IQ2_XXS | 8.15GB | false | Very low quality, uses SOTA techniques to be usable. |
+## 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.
+## Downloading using huggingface-cli
+<details>
+  <summary>Click to view download instructions</summary>
+First, make sure you have hugginface-cli installed:
+```
+pip install -U "huggingface_hub[cli]"
+```
+Then, you can target the specific file you want:
+```
+huggingface-cli download bartowski/Qwen_Qwen3-30B-A3B-GGUF --include "Qwen_Qwen3-30B-A3B-Q4_K_M.gguf" --local-dir ./
+```
+If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:
+```
+huggingface-cli download bartowski/Qwen_Qwen3-30B-A3B-GGUF --include "Qwen_Qwen3-30B-A3B-Q8_0/*" --local-dir ./
+```
+You can either specify a new local-dir (Qwen_Qwen3-30B-A3B-Q8_0) or download them all in place (./)
+</details>
+## ARM/AVX information
+Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass.
+Now, however, there is something called "online repacking" for weights. details in [this PR](https://github.com/ggerganov/llama.cpp/pull/9921). If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly.
+As of llama.cpp build [b4282](https://github.com/ggerganov/llama.cpp/releases/tag/b4282) you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0.
+Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to [this PR](https://github.com/ggerganov/llama.cpp/pull/10541) which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase.
+<details>
+  <summary>Click to view Q4_0_X_X information (deprecated</summary>
+I'm keeping this section to show the potential theoretical uplift in performance from using the Q4_0 with online repacking.
+<details>
+  <summary>Click to view benchmarks on an AVX2 system (EPYC7702)</summary>
+| model                          |       size |     params | backend    | threads |          test |                  t/s |  % (vs Q4_0)  |
+| ------------------------------ | ---------: | ---------: | ---------- | ------: | ------------: | -------------------: |-------------: |
+| qwen2 3B Q4_0                  |   1.70 GiB |     3.09 B | CPU        |      64 |         pp512 |        204.03 ± 1.03 |          100% |
+| qwen2 3B Q4_0                  |   1.70 GiB |     3.09 B | CPU        |      64 |        pp1024 |        282.92 ± 0.19 |          100% |
+| qwen2 3B Q4_0                  |   1.70 GiB |     3.09 B | CPU        |      64 |        pp2048 |        259.49 ± 0.44 |          100% |
+| qwen2 3B Q4_0                  |   1.70 GiB |     3.09 B | CPU        |      64 |         tg128 |         39.12 ± 0.27 |          100% |
+| qwen2 3B Q4_0                  |   1.70 GiB |     3.09 B | CPU        |      64 |         tg256 |         39.31 ± 0.69 |          100% |
+| qwen2 3B Q4_0                  |   1.70 GiB |     3.09 B | CPU        |      64 |         tg512 |         40.52 ± 0.03 |          100% |
+| qwen2 3B Q4_K_M                |   1.79 GiB |     3.09 B | CPU        |      64 |         pp512 |        301.02 ± 1.74 |          147% |
+| qwen2 3B Q4_K_M                |   1.79 GiB |     3.09 B | CPU        |      64 |        pp1024 |        287.23 ± 0.20 |          101% |
+| qwen2 3B Q4_K_M                |   1.79 GiB |     3.09 B | CPU        |      64 |        pp2048 |        262.77 ± 1.81 |          101% |
+| qwen2 3B Q4_K_M                |   1.79 GiB |     3.09 B | CPU        |      64 |         tg128 |         18.80 ± 0.99 |           48% |
+| qwen2 3B Q4_K_M                |   1.79 GiB |     3.09 B | CPU        |      64 |         tg256 |         24.46 ± 3.04 |           83% |
+| qwen2 3B Q4_K_M                |   1.79 GiB |     3.09 B | CPU        |      64 |         tg512 |         36.32 ± 3.59 |           90% |
+| qwen2 3B Q4_0_8_8              |   1.69 GiB |     3.09 B | CPU        |      64 |         pp512 |        271.71 ± 3.53 |          133% |
+| qwen2 3B Q4_0_8_8              |   1.69 GiB |     3.09 B | CPU        |      64 |        pp1024 |       279.86 ± 45.63 |          100% |
+| qwen2 3B Q4_0_8_8              |   1.69 GiB |     3.09 B | CPU        |      64 |        pp2048 |        320.77 ± 5.00 |          124% |
+| qwen2 3B Q4_0_8_8              |   1.69 GiB |     3.09 B | CPU        |      64 |         tg128 |         43.51 ± 0.05 |          111% |
+| qwen2 3B Q4_0_8_8              |   1.69 GiB |     3.09 B | CPU        |      64 |         tg256 |         43.35 ± 0.09 |          110% |
+| qwen2 3B Q4_0_8_8              |   1.69 GiB |     3.09 B | CPU        |      64 |         tg512 |         42.60 ± 0.31 |          105% |
+Q4_0_8_8 offers a nice bump to prompt processing and a small bump to text generation
+</details>
+</details>
+## Which file should I choose?
+<details>
+  <summary>Click here for details</summary>
+A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9)
+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.
+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](https://github.com/ggerganov/llama.cpp/wiki/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.
+Thank you to LM Studio for sponsoring my work.
+Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski