Qwen3-32B - DASH-Q 2-bit GGUF
llama.cpp GGUF files of Qwen/Qwen3-32B quantized with DASH-Q, at several 2-bit-class sizes. Every file uses only standard llama.cpp tensor types (no tensor above 4 bits) and loads in any recent llama.cpp build.
| File | Type | Size | Bits / weight |
|---|---|---|---|
Qwen3-32B-DASHQ-IQ2_XXS.gguf |
IQ2_XXS | 9.28 GB | 2.26 |
Qwen3-32B-DASHQ-IQ2_XS.gguf |
IQ2_XS | 10.80 GB | 2.63 |
Qwen3-32B-DASHQ-IQ2_M.gguf |
IQ2_M | 11.56 GB | 2.82 |
Qwen3-32B-DASHQ-Q2_K_XL.gguf |
Q2_K_XL | 12.63 GB | 3.08 |
Perplexity (lower is better)
| Type | Model | Size | WikiText-2 | C4 |
|---|---|---|---|---|
| IQ2_XXS | llama.cpp IQ2_XXS (imatrix) | 9.02 GB | 8.85 | 14.45 |
| IQ2_XXS | unsloth UD-IQ2_XXS | 9.28 GB | 8.74 | 14.09 |
| IQ2_XXS | DASH-Q IQ2_XXS | 9.28 GB | 8.11 | 13.58 |
| IQ2_XS | llama.cpp IQ2_XS (imatrix) | 9.95 GB | 8.14 | 13.31 |
| IQ2_XS | DASH-Q IQ2_XS | 10.80 GB | 7.37 | 12.33 |
| IQ2_M | llama.cpp IQ2_M (imatrix) | 11.36 GB | 7.53 | 12.67 |
| IQ2_M | unsloth UD-IQ2_M | 11.55 GB | 7.38 | 12.38 |
| IQ2_M | DASH-Q IQ2_M | 11.56 GB | 7.14 | 12.05 |
| Q2_K_XL | llama.cpp Q2_K (imatrix) | 12.34 GB | 14.85 | 22.03 |
| Q2_K_XL | unsloth UD-Q2_K_XL | 12.80 GB | 7.20 | 11.95 |
| Q2_K_XL | DASH-Q Q2_K_XL | 12.63 GB | 7.05 | 11.86 |
llama-perplexity, context 2048; WikiText-2 test, C4 validation (256 x 2048 tokens).
Usage
llama-cli -m Qwen3-32B-DASHQ-Q2_K_XL.gguf -ngl 99 -c 8192
License
Inherits the license of the base model (Qwen/Qwen3-32B).