granite-4.1-30b - DASH-Q 2-bit GGUF
llama.cpp GGUF files of ibm-granite/granite-4.1-30b 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 |
|---|---|---|---|
granite-4.1-30b-DASHQ-IQ2_XXS.gguf |
IQ2_XXS | 7.97 GB | 2.21 |
granite-4.1-30b-DASHQ-IQ2_XS.gguf |
IQ2_XS | 9.16 GB | 2.54 |
granite-4.1-30b-DASHQ-IQ2_M.gguf |
IQ2_M | 9.90 GB | 2.74 |
granite-4.1-30b-DASHQ-Q2_K_XL.gguf |
Q2_K_XL | 10.90 GB | 3.02 |
Perplexity (lower is better)
| Type | Model | Size | WikiText-2 | C4 |
|---|---|---|---|---|
| IQ2_XXS | llama.cpp IQ2_XXS (imatrix) | 7.78 GB | 9.41 | 15.62 |
| IQ2_XXS | unsloth UD-IQ2_XXS | 8.02 GB | 9.96 | 15.72 |
| IQ2_XXS | DASH-Q IQ2_XXS | 7.97 GB | 8.81 | 15.00 |
| IQ2_XS | llama.cpp IQ2_XS (imatrix) | 8.63 GB | 8.59 | 14.36 |
| IQ2_XS | DASH-Q IQ2_XS | 9.16 GB | 7.76 | 13.37 |
| IQ2_M | llama.cpp IQ2_M (imatrix) | 9.70 GB | 7.83 | 13.44 |
| IQ2_M | unsloth UD-IQ2_M | 9.94 GB | 7.83 | 13.12 |
| IQ2_M | DASH-Q IQ2_M | 9.90 GB | 7.46 | 12.88 |
| Q2_K_XL | llama.cpp Q2_K (imatrix) | 10.72 GB | 7.54 | 13.02 |
| Q2_K_XL | unsloth UD-Q2_K_XL | 10.98 GB | 7.61 | 12.82 |
| Q2_K_XL | DASH-Q Q2_K_XL | 10.90 GB | 7.39 | 12.77 |
llama-perplexity, context 2048; WikiText-2 test, C4 validation (256 x 2048 tokens).
Usage
llama-cli -m granite-4.1-30b-DASHQ-Q2_K_XL.gguf -ngl 99 -c 8192
License
Inherits the license of the base model (ibm-granite/granite-4.1-30b).