About
weighted/imatrix quants of https://huggingface.co/nbeerbower/Gemma4-Gutenberg-26B-A4B
Benchmarked on CaliperBench — a creative-writing benchmark scoring prose craft, roleplay and willingness rather than general intelligence. See this model's scores: caliperbench.com/m/gemma4-gutenberg-26b-a4b.
Note: Gemma4-Gutenberg-26B-A4B is published as a LoRA adapter, not a full model. These GGUFs are that adapter merged onto its base google/gemma-4-26B-A4B-it (the text LM extracted from the multimodal base), then quantized.
imatrix (importance matrix) computed over bartowski calibration_datav3. Weighted quants spend precision where it matters most — the low bit-rates (IQ1–IQ4) are meaningfully better than same-size static quants; at Q5/Q6 the difference is negligible.
Static (non-imatrix) quants of this model are at spiritfather/Gemma4-Gutenberg-26B-A4B-GGUF.
Usage
If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files.
Provided Quants
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
| Link | Type | Size/GB | Notes |
|---|---|---|---|
| GGUF | IQ1_S | 8.3 | for the desperate |
| GGUF | IQ1_M | 8.7 | mostly desperate |
| GGUF | IQ2_XXS | 9.3 | lower quality |
| GGUF | IQ2_XS | 9.8 | |
| GGUF | IQ2_S | 9.9 | |
| GGUF | IQ2_M | 10.4 | |
| GGUF | Q2_K_S | 10.6 | |
| GGUF | Q2_K | 10.6 | |
| GGUF | IQ3_XXS | 11.3 | lower quality |
| GGUF | IQ3_XS | 11.6 | |
| GGUF | Q3_K_S | 12.2 | |
| GGUF | IQ3_S | 12.2 | |
| GGUF | IQ3_M | 12.4 | |
| GGUF | Q3_K_M | 13.3 | IQ3_S probably better |
| GGUF | Q3_K_L | 13.8 | |
| GGUF | IQ4_XS | 13.9 | |
| GGUF | Q4_0 | 14.5 | |
| GGUF | Q4_K_S | 15.5 | optimal size/speed/quality |
| GGUF | Q4_K_M | 16.8 | fast, recommended |
| GGUF | Q4_1 | 16.0 | |
| GGUF | Q5_K_S | 18.0 | |
| GGUF | Q5_K_M | 19.1 | |
| GGUF | Q6_K | 22.6 | practically like static Q6_K |