elichen-skymizer
GAP-models
- Parameters
- 30.7B
- Max context length
- 262,144
- Architecture
- gemma4
- llama.cpp
- b8637+
About GAP-models
GAP-models is a model published in GGUF format. GGUF files are published by elichen-skymizer/GAP-models. It has about 30.7B parameters, the gemma4 architecture, a maximum context length of 262,144 tokens (values written in the GGUF).
There are 14 quantizations, with file sizes from 2.48 GB (IQ4_XS) to 71.1 GB (BF16) (actual sizes on Hugging Face).
Running Q4_K_M (2.74 GB) with an 8,192-token context needs an estimated 3.87 GB of memory, and an estimated 4.63 GB at 32,768 tokens. For 8,192 tokens, a GPU with 4 GB or more of VRAM (such as Arc A310, Arc A370M), or a Mac with 4 GB or more of unified memory, is a rough guide (macOS caps how much memory the GPU can use, so leave some headroom). The smallest, IQ4_XS (2.48 GB), needs an estimated 3.60 GB at 8,192 tokens.
File size and estimated memory per quantization
- Up to 8 GB
- Up to 16 GB
- Up to 24 GB
- Up to 48 GB
- Up to 128 GB
- More
File sizes are the actual values on Hugging Face. Memory is estimated for each context length (KV cache f16). The formula is in How memory is calculated.
All memory figures are Estimated (KV cache f16). It should run on a setup with at least this much VRAM (or Mac unified memory). Click a value to see the devices and models that fit that amount (Find by GPU / Mac).
Using image input: This model supports image input. To use images, an mmproj file is loaded as well.
About the "quality guide" (how far a quantization drifts from the original)
For each quantization type, how far the output drifts from the original model, grouped into six levels from values llama.cpp measured on Llama-3-8B.
The value is mean KLD (the difference between the original and quantized model in next-token probabilities); smaller means closer to the original. Quantizations made with an imatrix (importance data) come out slightly smaller. Levels use the value without an imatrix (or with one, for types that only have that).
These values were not measured on this model. Larger models and MoE models tend to drift less with the same type. Distributor-specific mixed types (such as UD-Q4_K_XL) cannot be matched by type name alone, so they show "-".
| Quantization | Quality guide | KLD (no imatrix) | KLD (imatrix) | Size (Llama-3-8B) |
|---|---|---|---|---|
| Q8_0 | Nearly identical | 0.0014 | - | 7.96 GiB |
| Q6_K | Very small | 0.0055 | - | 6.14 GiB |
| Q5_K_M | Very small | 0.0108 | - | 5.33 GiB |
| Q5_K_S | Small | 0.0166 | - | 5.21 GiB |
| Q5_1 | Small | 0.0180 | - | 5.65 GiB |
| Q5_0 | Small | 0.0222 | - | 5.21 GiB |
| Q4_K_M | Small | 0.0313 | 0.0282 | 4.58 GiB |
| Q4_K_S | Medium | 0.0431 | 0.0320 | 4.37 GiB |
| IQ4_NL | Small | - | 0.0357 | 4.35 GiB |
| IQ4_XS | Small | - | 0.0363 | 4.14 GiB |
| Q4_1 | Medium | 0.0717 | - | 4.78 GiB |
| Q4_0 | Medium | 0.0719 | - | 4.34 GiB |
| Q3_K_L | Medium | 0.0872 | 0.0731 | 4.03 GiB |
| Q3_K_M | Medium | 0.1019 | 0.0844 | 3.74 GiB |
| IQ3_M | Medium | - | 0.1025 | 3.53 GiB |
| IQ3_S | Medium | - | 0.1113 | 3.42 GiB |
| IQ3_XS | Large | - | 0.1387 | 3.28 GiB |
| IQ3_XXS | Large | - | 0.1836 | 3.05 GiB |
| Q3_K_S | Large | 0.2282 | 0.1998 | 3.41 GiB |
| IQ2_M | Large | - | 0.3260 | 2.74 GiB |
| Q2_K | Very large | 0.4451 | 0.3322 | 2.96 GiB |
| Q2_K_S | Very large | - | 0.4034 | 2.96 GiB |
| IQ2_S | Very large | - | 0.4392 | 2.56 GiB |
| IQ2_XS | Very large | - | 0.5463 | 2.43 GiB |
| IQ2_XXS | Very large | - | 0.8120 | 2.24 GiB |
| IQ1_M | Very large | - | 1.3931 | 2.01 GiB |
| IQ1_S | Very large | - | 2.2546 | 1.88 GiB |
Source: llama.cpp (MIT), measured on Llama-3-8B against FP16, Wikitext-2, revision f364eb6f. Original table · Reading and choosing quantizations
Last checked: Oct 11, 2026
- Architecture
- gemma4
- llama.cpp b8637 and later
- Parameters
- 30.7B
- Max context length
- 262,144
| Quantization | File size Measured | Quality guide | 4,096 | 8,192 | 16,384 | 32,768 | 65,536 | 131,072 | 262,144 max |
|---|---|---|---|---|---|---|---|---|---|
| IQ4_XS Qwen3.5-4B-IQ4_XS | 2.48 GB | Small | 3.48 GB | 3.60 GB | 3.86 GB | 4.37 GB | 5.39 GB | 7.43 GB | 11.5 GB |
| IQ4_NL Qwen3.5-4B-IQ4_NL | 2.58 GB | Small | 3.58 GB | 3.71 GB | 3.96 GB | 4.47 GB | 5.49 GB | 7.53 GB | 11.6 GB |
| Q4_0 Qwen3.5-4B-Q4_0 | 2.58 GB | Medium | 3.58 GB | 3.71 GB | 3.97 GB | 4.47 GB | 5.49 GB | 7.53 GB | 11.6 GB |
| Q4_K_S Qwen3.5-4B-Q4_K_S | 2.59 GB | Medium | 3.59 GB | 3.72 GB | 3.97 GB | 4.48 GB | 5.50 GB | 7.54 GB | 11.6 GB |
| IQ4_XS Qwen_Qwen3.5-4B-IQ4_XS | 2.67 GB | Small | 3.67 GB | 3.80 GB | 4.05 GB | 4.56 GB | 5.58 GB | 7.62 GB | 11.7 GB |
| Q4_K_M Qwen3.5-4B-Q4_K_M | 2.74 GB | Small | 3.74 GB | 3.87 GB | 4.12 GB | 4.63 GB | 5.65 GB | 7.69 GB | 11.8 GB |
| IQ4_NL Qwen_Qwen3.5-4B-IQ4_NL | 2.78 GB | Small | 3.78 GB | 3.90 GB | 4.16 GB | 4.67 GB | 5.69 GB | 7.72 GB | 11.8 GB |
| Q4_0 Qwen_Qwen3.5-4B-Q4_0 | 2.78 GB | Medium | 3.78 GB | 3.90 GB | 4.16 GB | 4.67 GB | 5.69 GB | 7.73 GB | 11.8 GB |
| Q4_1 Qwen3.5-4B-Q4_1 | 2.78 GB | Medium | 3.78 GB | 3.91 GB | 4.17 GB | 4.68 GB | 5.70 GB | 7.73 GB | 11.8 GB |
| Q4_K_S Qwen_Qwen3.5-4B-Q4_K_S | 2.85 GB | Medium | 3.85 GB | 3.97 GB | 4.23 GB | 4.74 GB | 5.76 GB | 7.80 GB | 11.9 GB |
| UD-Q4_K_XL Qwen3.5-4B-UD-Q4_K_XL | 2.91 GB | - | 3.91 GB | 4.04 GB | 4.29 GB | 4.80 GB | 5.82 GB | 7.86 GB | 11.9 GB |
| Q4_1 Qwen_Qwen3.5-4B-Q4_1 | 2.98 GB | Medium | 3.98 GB | 4.11 GB | 4.36 GB | 4.87 GB | 5.89 GB | 7.93 GB | 12.0 GB |
| Q4_K_M Qwen_Qwen3.5-4B-Q4_K_M | 3.01 GB | Small | 4.01 GB | 4.14 GB | 4.40 GB | 4.90 GB | 5.92 GB | 7.96 GB | 12.0 GB |
| UD-Q3_K_XL gemma-4-E4B-it-UD-Q3_K_XL | 4.59 GB | - | 5.59 GB | 5.72 GB | 5.97 GB | 6.48 GB | 7.50 GB | 9.54 GB | 13.6 GB |
| IQ4_XS gemma-4-E4B-it-IQ4_XS | 4.72 GB | Small | 5.72 GB | 5.84 GB | 6.10 GB | 6.61 GB | 7.63 GB | 9.66 GB | 13.7 GB |
| IQ4_NL gemma-4-E4B-it-IQ4_NL | 4.84 GB | Small | 5.84 GB | 5.96 GB | 6.22 GB | 6.73 GB | 7.75 GB | 9.79 GB | 13.9 GB |
| Q4_0 gemma-4-E4B-it-Q4_0 | 4.84 GB | Medium | 5.84 GB | 5.96 GB | 6.22 GB | 6.73 GB | 7.75 GB | 9.79 GB | 13.9 GB |
| Q4_K_S gemma-4-E4B-it-Q4_K_S | 4.84 GB | Medium | 5.84 GB | 5.97 GB | 6.23 GB | 6.74 GB | 7.76 GB | 9.79 GB | 13.9 GB |
| Q4_K_M gemma-4-E4B-it-Q4_K_M | 4.98 GB | Small | 5.98 GB | 6.10 GB | 6.36 GB | 6.87 GB | 7.89 GB | 9.93 GB | 14.0 GB |
| Q4_1 gemma-4-E4B-it-Q4_1 | 5.07 GB | Medium | 6.07 GB | 6.20 GB | 6.46 GB | 6.97 GB | 7.99 GB | 10.0 GB | 14.1 GB |
| IQ4_XS google_gemma-4-E4B-it-IQ4_XS | 5.11 GB | Small | 6.11 GB | 6.24 GB | 6.49 GB | 7.00 GB | 8.02 GB | 10.1 GB | 14.1 GB |
| UD-Q4_K_XL gemma-4-E4B-it-UD-Q4_K_XL | 5.13 GB | - | 6.13 GB | 6.25 GB | 6.51 GB | 7.02 GB | 8.04 GB | 10.1 GB | 14.2 GB |
| Q4_0 gemma-4-E4B_q4_0-it | 5.15 GB | Medium | 6.15 GB | 6.28 GB | 6.54 GB | 7.05 GB | 8.07 GB | 10.1 GB | 14.2 GB |
| Q4_0 google_gemma-4-E4B-it-Q4_0 | 5.22 GB | Medium | 6.22 GB | 6.35 GB | 6.61 GB | 7.12 GB | 8.13 GB | 10.2 GB | 14.3 GB |
| IQ4_NL google_gemma-4-E4B-it-IQ4_NL | 5.23 GB | Small | 6.23 GB | 6.36 GB | 6.61 GB | 7.12 GB | 8.14 GB | 10.2 GB | 14.3 GB |
| Q4_K_S google_gemma-4-E4B-it-Q4_K_S | 5.24 GB | Medium | 6.24 GB | 6.36 GB | 6.62 GB | 7.13 GB | 8.15 GB | 10.2 GB | 14.3 GB |
| IQ4_XS zai-org_GLM-4.6V-Flash-IQ4_XS | 5.26 GB | Small | 6.26 GB | 6.39 GB | 6.64 GB | 7.15 GB | 8.17 GB | 10.2 GB | 14.3 GB |
| IQ4_XS GLM-4.6V-Flash-IQ4_XS | 5.28 GB | Small | 6.28 GB | 6.41 GB | 6.66 GB | 7.17 GB | 8.19 GB | 10.2 GB | 14.3 GB |
| Q4_K_M google_gemma-4-E4B-it-Q4_K_M | 5.41 GB | Small | 6.40 GB | 6.53 GB | 6.79 GB | 7.30 GB | 8.32 GB | 10.4 GB | 14.4 GB |
| Q4_1 google_gemma-4-E4B-it-Q4_1 | 5.46 GB | Medium | 6.46 GB | 6.59 GB | 6.84 GB | 7.35 GB | 8.37 GB | 10.4 GB | 14.5 GB |
| IQ4_NL zai-org_GLM-4.6V-Flash-IQ4_NL | 5.47 GB | Small | 6.46 GB | 6.59 GB | 6.85 GB | 7.36 GB | 8.38 GB | 10.4 GB | 14.5 GB |
| IQ4_NL GLM-4.6V-Flash-IQ4_NL | 5.47 GB | Small | 6.46 GB | 6.59 GB | 6.85 GB | 7.36 GB | 8.38 GB | 10.4 GB | 14.5 GB |
| Q4_0 zai-org_GLM-4.6V-Flash-Q4_0 | 5.48 GB | Medium | 6.48 GB | 6.60 GB | 6.86 GB | 7.37 GB | 8.39 GB | 10.4 GB | 14.5 GB |
| Q4_0 GLM-4.6V-Flash-Q4_0 | 5.48 GB | Medium | 6.48 GB | 6.60 GB | 6.86 GB | 7.37 GB | 8.39 GB | 10.4 GB | 14.5 GB |
| Q3_K_XL zai-org_GLM-4.6V-Flash-Q3_K_XL | 5.74 GB | - | 6.74 GB | 6.87 GB | 7.12 GB | 7.63 GB | 8.65 GB | 10.7 GB | 14.8 GB |
| Q4_K_S zai-org_GLM-4.6V-Flash-Q4_K_S | 5.76 GB | Medium | 6.76 GB | 6.89 GB | 7.14 GB | 7.65 GB | 8.67 GB | 10.7 GB | 14.8 GB |
| Q4_K_S GLM-4.6V-Flash-Q4_K_S | 5.76 GB | Medium | 6.76 GB | 6.89 GB | 7.14 GB | 7.65 GB | 8.67 GB | 10.7 GB | 14.8 GB |
| Q3_K_XL google_gemma-4-E4B-it-Q3_K_XL | 5.88 GB | - | 6.88 GB | 7.01 GB | 7.26 GB | 7.77 GB | 8.79 GB | 10.8 GB | 14.9 GB |
| Q4_1 zai-org_GLM-4.6V-Flash-Q4_1 | 6.01 GB | Medium | 7.01 GB | 7.14 GB | 7.39 GB | 7.90 GB | 8.92 GB | 11.0 GB | 15.0 GB |
| Q4_1 GLM-4.6V-Flash-Q4_1 | 6.01 GB | Medium | 7.01 GB | 7.14 GB | 7.39 GB | 7.90 GB | 8.92 GB | 11.0 GB | 15.0 GB |
| Q4_K_M zai-org_GLM-4.6V-Flash-Q4_K_M | 6.17 GB | Small | 7.17 GB | 7.29 GB | 7.55 GB | 8.06 GB | 9.08 GB | 11.1 GB | 15.2 GB |
| Q4_K_M GLM-4.6V-Flash-Q4_K_M | 6.17 GB | Small | 7.17 GB | 7.29 GB | 7.55 GB | 8.06 GB | 9.08 GB | 11.1 GB | 15.2 GB |
| UD-Q4_K_XL GLM-4.6V-Flash-UD-Q4_K_XL | 6.19 GB | - | 7.19 GB | 7.31 GB | 7.57 GB | 8.08 GB | 9.10 GB | 11.1 GB | 15.2 GB |
| BF16 Qwen_Qwen3.5-4B-bf16 | 8.67 GB | Original precision | 9.67 GB | 9.79 GB | 10.0 GB | 10.6 GB | 11.6 GB | 13.6 GB | 17.7 GB |
| IQ4_XS Kimi-VL-A3B-Instruct.i1-IQ4_XS | 8.75 GB | Small | 9.75 GB | 9.87 GB | 10.1 GB | 10.6 GB | 11.7 GB | 13.7 GB | 17.8 GB |
| IQ4_XS Kimi-VL-A3B-Thinking-2506.i1-IQ4_XS | 8.75 GB | Small | 9.75 GB | 9.87 GB | 10.1 GB | 10.6 GB | 11.7 GB | 13.7 GB | 17.8 GB |
| IQ4_XS Kimi-VL-A3B-Instruct-IQ4_XS | 8.82 GB | Small | 9.82 GB | 9.95 GB | 10.2 GB | 10.7 GB | 11.7 GB | 13.8 GB | 17.8 GB |
| IQ4_XS Kimi-VL-A3B-Instruct.IQ4_XS | 8.82 GB | Small | 9.82 GB | 9.95 GB | 10.2 GB | 10.7 GB | 11.7 GB | 13.8 GB | 17.8 GB |
| Q4_0 Kimi-VL-A3B-Thinking-2506-Q4_0 | 9.08 GB | Medium | 10.1 GB | 10.2 GB | 10.5 GB | 11.0 GB | 12.0 GB | 14.0 GB | 18.1 GB |
| Q4_0 Kimi-VL-A3B-Instruct-Q4_0 | 9.08 GB | Medium | 10.1 GB | 10.2 GB | 10.5 GB | 11.0 GB | 12.0 GB | 14.0 GB | 18.1 GB |
| IQ4_NL Kimi-VL-A3B-Instruct.i1-IQ4_NL | 9.08 GB | Small | 10.1 GB | 10.2 GB | 10.5 GB | 11.0 GB | 12.0 GB | 14.0 GB | 18.1 GB |
| IQ4_NL Kimi-VL-A3B-Thinking-2506.i1-IQ4_NL | 9.08 GB | Small | 10.1 GB | 10.2 GB | 10.5 GB | 11.0 GB | 12.0 GB | 14.0 GB | 18.1 GB |
| Q4_0 Kimi-VL-A3B-Instruct.i1-Q4_0 | 9.11 GB | Medium | 10.1 GB | 10.2 GB | 10.5 GB | 11.0 GB | 12.0 GB | 14.1 GB | 18.1 GB |
| Q4_0 Kimi-VL-A3B-Thinking-2506.i1-Q4_0 | 9.11 GB | Medium | 10.1 GB | 10.2 GB | 10.5 GB | 11.0 GB | 12.0 GB | 14.1 GB | 18.1 GB |
| IQ4_NL Kimi-VL-A3B-Instruct-IQ4_NL | 9.16 GB | Small | 10.2 GB | 10.3 GB | 10.5 GB | 11.0 GB | 12.1 GB | 14.1 GB | 18.2 GB |
| Q4_K_S Kimi-VL-A3B-Thinking-2506-Q4_K_S | 9.71 GB | Medium | 10.7 GB | 10.8 GB | 11.1 GB | 11.6 GB | 12.6 GB | 14.7 GB | 18.7 GB |
| Q4_K_S Kimi-VL-A3B-Instruct-Q4_K_S | 9.71 GB | Medium | 10.7 GB | 10.8 GB | 11.1 GB | 11.6 GB | 12.6 GB | 14.7 GB | 18.7 GB |
| Q4_K_S Kimi-VL-A3B-Instruct.Q4_K_S | 9.71 GB | Medium | 10.7 GB | 10.8 GB | 11.1 GB | 11.6 GB | 12.6 GB | 14.7 GB | 18.7 GB |
| Q4_K_S Kimi-VL-A3B-Thinking-2506.Q4_K_S | 9.71 GB | Medium | 10.7 GB | 10.8 GB | 11.1 GB | 11.6 GB | 12.6 GB | 14.7 GB | 18.7 GB |
| Q4_K_S Kimi-VL-A3B-Thinking-2506.i1-Q4_K_S | 9.71 GB | Medium | 10.7 GB | 10.8 GB | 11.1 GB | 11.6 GB | 12.6 GB | 14.7 GB | 18.7 GB |
| Q4_1 Kimi-VL-A3B-Thinking-2506-Q4_1 | 10.1 GB | Medium | 11.1 GB | 11.2 GB | 11.4 GB | 12.0 GB | 13.0 GB | 15.0 GB | 19.1 GB |
| Q4_1 Kimi-VL-A3B-Instruct-Q4_1 | 10.1 GB | Medium | 11.1 GB | 11.2 GB | 11.4 GB | 12.0 GB | 13.0 GB | 15.0 GB | 19.1 GB |
| Q4_1 Kimi-VL-A3B-Instruct.i1-Q4_1 | 10.1 GB | Medium | 11.1 GB | 11.2 GB | 11.4 GB | 12.0 GB | 13.0 GB | 15.0 GB | 19.1 GB |
| Q4_1 Kimi-VL-A3B-Thinking-2506.i1-Q4_1 | 10.1 GB | Medium | 11.1 GB | 11.2 GB | 11.4 GB | 12.0 GB | 13.0 GB | 15.0 GB | 19.1 GB |
| Q4_K_M Kimi-VL-A3B-Thinking-2506-Q4_K_M | 10.5 GB | Small | 11.5 GB | 11.7 GB | 11.9 GB | 12.4 GB | 13.5 GB | 15.5 GB | 19.6 GB |
| Q4_K_M Kimi-VL-A3B-Instruct-Q4_K_M | 10.5 GB | Small | 11.5 GB | 11.7 GB | 11.9 GB | 12.4 GB | 13.5 GB | 15.5 GB | 19.6 GB |
| Q4_K_M Kimi-VL-A3B-Instruct.Q4_K_M | 10.5 GB | Small | 11.5 GB | 11.7 GB | 11.9 GB | 12.4 GB | 13.5 GB | 15.5 GB | 19.6 GB |
| Q4_K_M Kimi-VL-A3B-Thinking-2506.Q4_K_M | 10.5 GB | Small | 11.5 GB | 11.7 GB | 11.9 GB | 12.4 GB | 13.5 GB | 15.5 GB | 19.6 GB |
| Q4_K_M Kimi-VL-A3B-Instruct.i1-Q4_K_M | 10.5 GB | Small | 11.5 GB | 11.7 GB | 11.9 GB | 12.4 GB | 13.5 GB | 15.5 GB | 19.6 GB |
| Q4_K_M Kimi-VL-A3B-Thinking-2506.i1-Q4_K_M | 10.5 GB | Small | 11.5 GB | 11.7 GB | 11.9 GB | 12.4 GB | 13.5 GB | 15.5 GB | 19.6 GB |
| UD-Q3_K_XL Muse-Glimmer-30B-UD-Q3_K_XL | 13.4 GB | - | 14.4 GB | 14.5 GB | 14.7 GB | 15.3 GB | 16.3 GB | 18.3 GB | 22.4 GB |
| Q3_K_L | 14.7 GB | Medium | 15.7 GB | 15.8 GB | 16.1 GB | 16.6 GB | 17.6 GB | 19.6 GB | 23.7 GB |
| Q3_K_XL OpenGVLab_InternVL3_5-30B-A3B-Q3_K_XL | 14.9 GB | - | 15.9 GB | 16.0 GB | 16.2 GB | 16.7 GB | 17.8 GB | 19.8 GB | 23.9 GB |
| BF16 google_gemma-4-E4B-it-bf16 | 15.1 GB | Original precision | 16.1 GB | 16.2 GB | 16.4 GB | 16.9 GB | 18.0 GB | 20.0 GB | 24.1 GB |
| IQ4_XS Muse-Glimmer-30B.i1-IQ4_XS | 15.2 GB | Small | 16.2 GB | 16.3 GB | 16.6 GB | 17.1 GB | 18.1 GB | 20.2 GB | 24.2 GB |
| IQ4_XS Muse-Glimmer-30B.IQ4_XS | 15.3 GB | Small | 16.3 GB | 16.5 GB | 16.7 GB | 17.2 GB | 18.3 GB | 20.3 GB | 24.4 GB |
| UD-Q3_K_XL gemma-4-31B-it-UD-Q3_K_XL | 15.4 GB | - | 16.4 GB | 16.5 GB | 16.8 GB | 17.3 GB | 18.3 GB | 20.3 GB | 24.4 GB |
| IQ4_XS Muse-Glimmer-30B-IQ4_XS | 15.4 GB | Small | 16.4 GB | 16.6 GB | 16.8 GB | 17.3 GB | 18.3 GB | 20.4 GB | 24.5 GB |
| UD-Q4_K_XL Muse-Glimmer-30B-UD-Q4_K_XL | 15.9 GB | - | 16.9 GB | 17.0 GB | 17.3 GB | 17.8 GB | 18.8 GB | 20.8 GB | 24.9 GB |
| Q4_0 Muse-Glimmer-30B.i1-Q4_0 | 16.1 GB | Medium | 17.1 GB | 17.2 GB | 17.5 GB | 18.0 GB | 19.0 GB | 21.0 GB | 25.1 GB |
| Q4_K_S Muse-Glimmer-30B.Q4_K_S | 16.1 GB | Medium | 17.1 GB | 17.3 GB | 17.5 GB | 18.0 GB | 19.0 GB | 21.1 GB | 25.2 GB |
| Q4_K_S Muse-Glimmer-30B.i1-Q4_K_S | 16.1 GB | Medium | 17.1 GB | 17.3 GB | 17.5 GB | 18.0 GB | 19.0 GB | 21.1 GB | 25.2 GB |
| IQ4_NL Muse-Glimmer-30B-IQ4_NL | 16.2 GB | Small | 17.2 GB | 17.4 GB | 17.6 GB | 18.1 GB | 19.1 GB | 21.2 GB | 25.3 GB |
| Q4_0 Muse-Glimmer-30B-Q4_0 | 16.3 GB | Medium | 17.3 GB | 17.4 GB | 17.6 GB | 18.2 GB | 19.2 GB | 21.2 GB | 25.3 GB |
| Q4_K_S Muse-Glimmer-30B-Q4_K_S | 16.3 GB | Medium | 17.3 GB | 17.4 GB | 17.7 GB | 18.2 GB | 19.2 GB | 21.3 GB | 25.3 GB |
| IQ4_XS InternVL3_5-30B-A3B.i1-IQ4_XS | 16.4 GB | Small | 17.4 GB | 17.5 GB | 17.8 GB | 18.3 GB | 19.3 GB | 21.3 GB | 25.4 GB |
| IQ4_XS gemma-4-31B-it-IQ4_XS | 16.4 GB | Small | 17.4 GB | 17.5 GB | 17.8 GB | 18.3 GB | 19.3 GB | 21.3 GB | 25.4 GB |
| IQ4_XS OpenGVLab_InternVL3_5-30B-A3B-IQ4_XS | 16.5 GB | Small | 17.5 GB | 17.6 GB | 17.8 GB | 18.3 GB | 19.4 GB | 21.4 GB | 25.5 GB |
| IQ4_XS InternVL3_5-30B-A3B.IQ4_XS | 16.6 GB | Small | 17.6 GB | 17.7 GB | 17.9 GB | 18.4 GB | 19.5 GB | 21.5 GB | 25.6 GB |
| Q4_K_M Muse-Glimmer-30B.Q4_K_M | 16.9 GB | Small | 17.9 GB | 18.1 GB | 18.3 GB | 18.8 GB | 19.8 GB | 21.9 GB | 26.0 GB |
| Q4_K_M Muse-Glimmer-30B.i1-Q4_K_M | 16.9 GB | Small | 17.9 GB | 18.1 GB | 18.3 GB | 18.8 GB | 19.8 GB | 21.9 GB | 26.0 GB |
| Q3_K_XL google_gemma-4-31B-it-Q3_K_XL | 17.2 GB | - | 18.2 GB | 18.3 GB | 18.5 GB | 19.0 GB | 20.1 GB | 22.1 GB | 26.2 GB |
| IQ4_XS google_gemma-4-31B-it-IQ4_XS | 17.2 GB | Small | 18.2 GB | 18.3 GB | 18.5 GB | 19.0 GB | 20.1 GB | 22.1 GB | 26.2 GB |
| IQ4_NL gemma-4-31B-it-IQ4_NL | 17.3 GB | Small | 18.3 GB | 18.4 GB | 18.7 GB | 19.2 GB | 20.2 GB | 22.2 GB | 26.3 GB |
| Q4_K_M Muse-Glimmer-30B-Q4_K_M | 17.3 GB | Small | 18.3 GB | 18.4 GB | 18.7 GB | 19.2 GB | 20.2 GB | 22.3 GB | 26.3 GB |
| Q4_0 gemma-4-31B-it-Q4_0 | 17.3 GB | Medium | 18.3 GB | 18.5 GB | 18.7 GB | 19.2 GB | 20.2 GB | 22.3 GB | 26.4 GB |
| Q4_0 InternVL3_5-30B-A3B.i1-Q4_0 | 17.4 GB | Medium | 18.4 GB | 18.5 GB | 18.8 GB | 19.3 GB | 20.3 GB | 22.3 GB | 26.4 GB |
| IQ4_NL OpenGVLab_InternVL3_5-30B-A3B-IQ4_NL | 17.4 GB | Small | 18.4 GB | 18.5 GB | 18.8 GB | 19.3 GB | 20.3 GB | 22.3 GB | 26.4 GB |
| Q4_K_S gemma-4-31B-it-Q4_K_S | 17.4 GB | Medium | 18.4 GB | 18.5 GB | 18.8 GB | 19.3 GB | 20.3 GB | 22.3 GB | 26.4 GB |
| Q4_K_S InternVL3_5-30B-A3B.Q4_K_S | 17.5 GB | Medium | 18.5 GB | 18.6 GB | 18.8 GB | 19.3 GB | 20.4 GB | 22.4 GB | 26.5 GB |
| Q4_K_S InternVL3_5-30B-A3B.i1-Q4_K_S | 17.5 GB | Medium | 18.5 GB | 18.6 GB | 18.8 GB | 19.3 GB | 20.4 GB | 22.4 GB | 26.5 GB |
| Q4_0 OpenGVLab_InternVL3_5-30B-A3B-Q4_0 | 17.6 GB | Medium | 18.6 GB | 18.8 GB | 19.0 GB | 19.5 GB | 20.5 GB | 22.6 GB | 26.7 GB |
| Q4_0 gemma-4-31B_q4_0-it | 17.7 GB | Medium | 18.7 GB | 18.8 GB | 19.0 GB | 19.5 GB | 20.6 GB | 22.6 GB | 26.7 GB |
| Q4_1 Muse-Glimmer-30B.i1-Q4_1 | 17.7 GB | Medium | 18.7 GB | 18.8 GB | 19.1 GB | 19.6 GB | 20.6 GB | 22.6 GB | 26.7 GB |
| Q4_1 Muse-Glimmer-30B-Q4_1 | 17.8 GB | Medium | 18.8 GB | 19.0 GB | 19.2 GB | 19.7 GB | 20.7 GB | 22.8 GB | 26.9 GB |
| Q4_K_S OpenGVLab_InternVL3_5-30B-A3B-Q4_K_S | 18.0 GB | Medium | 19.0 GB | 19.1 GB | 19.4 GB | 19.9 GB | 20.9 GB | 22.9 GB | 27.0 GB |
| Q4_0 google_gemma-4-31B-it-Q4_0 | 18.1 GB | Medium | 19.1 GB | 19.2 GB | 19.5 GB | 20.0 GB | 21.0 GB | 23.0 GB | 27.1 GB |
| Q4_K_S google_gemma-4-31B-it-Q4_K_S | 18.2 GB | Medium | 19.2 GB | 19.3 GB | 19.6 GB | 20.1 GB | 21.1 GB | 23.1 GB | 27.2 GB |
| Q4_K_M gemma-4-31B-it-Q4_K_M | 18.3 GB | Small | 19.3 GB | 19.5 GB | 19.7 GB | 20.2 GB | 21.2 GB | 23.3 GB | 27.3 GB |
| Q4_K_M InternVL3_5-30B-A3B.Q4_K_M | 18.6 GB | Small | 19.6 GB | 19.7 GB | 19.9 GB | 20.4 GB | 21.5 GB | 23.5 GB | 27.6 GB |
| Q4_K_M InternVL3_5-30B-A3B.i1-Q4_K_M | 18.6 GB | Small | 19.6 GB | 19.7 GB | 19.9 GB | 20.4 GB | 21.5 GB | 23.5 GB | 27.6 GB |
| Q4_K_M OpenGVLab_InternVL3_5-30B-A3B-Q4_K_M | 18.6 GB | Small | 19.6 GB | 19.8 GB | 20.0 GB | 20.5 GB | 21.5 GB | 23.6 GB | 27.7 GB |
| BF16 zai-org_GLM-4.6V-Flash-bf16 | 18.8 GB | Original precision | 19.8 GB | 19.9 GB | 20.2 GB | 20.7 GB | 21.7 GB | 23.8 GB | 27.8 GB |
| UD-Q4_K_XL gemma-4-31B-it-UD-Q4_K_XL | 18.8 GB | - | 19.8 GB | 20.0 GB | 20.2 GB | 20.7 GB | 21.7 GB | 23.8 GB | 27.8 GB |
| Q4_1 gemma-4-31B-it-Q4_1 | 19.1 GB | Medium | 20.1 GB | 20.2 GB | 20.5 GB | 21.0 GB | 22.0 GB | 24.1 GB | 28.1 GB |
| Q4_1 InternVL3_5-30B-A3B.i1-Q4_1 | 19.2 GB | Medium | 20.2 GB | 20.3 GB | 20.6 GB | 21.1 GB | 22.1 GB | 24.1 GB | 28.2 GB |
| Q4_1 OpenGVLab_InternVL3_5-30B-A3B-Q4_1 | 19.2 GB | Medium | 20.2 GB | 20.3 GB | 20.6 GB | 21.1 GB | 22.1 GB | 24.2 GB | 28.2 GB |
| Q4_K_M google_gemma-4-31B-it-Q4_K_M | 19.6 GB | Small | 20.6 GB | 20.7 GB | 21.0 GB | 21.5 GB | 22.5 GB | 24.5 GB | 28.6 GB |
| IQ4_XS Qwen_Qwen3.6-35B-A3B-IQ4_XS | 19.7 GB | Small | 20.7 GB | 20.8 GB | 21.1 GB | 21.6 GB | 22.6 GB | 24.6 GB | 28.7 GB |
| Q4_1 google_gemma-4-31B-it-Q4_1 | 19.8 GB | Medium | 20.8 GB | 20.9 GB | 21.1 GB | 21.7 GB | 22.7 GB | 24.7 GB | 28.8 GB |
| IQ4_NL Qwen_Qwen3.6-35B-A3B-IQ4_NL | 20.8 GB | Small | 21.8 GB | 21.9 GB | 22.1 GB | 22.6 GB | 23.7 GB | 25.7 GB | 29.8 GB |
| Q4_0 Qwen_Qwen3.6-35B-A3B-Q4_0 | 20.8 GB | Medium | 21.8 GB | 22.0 GB | 22.2 GB | 22.7 GB | 23.7 GB | 25.8 GB | 29.9 GB |
| UD-Q4_K_S | 20.9 GB | - | 21.9 GB | 22.0 GB | 22.3 GB | 22.8 GB | 23.8 GB | 25.8 GB | 29.9 GB |
| Q4_K_S Qwen_Qwen3.6-35B-A3B-Q4_K_S | 21.5 GB | Medium | 22.5 GB | 22.6 GB | 22.9 GB | 23.4 GB | 24.4 GB | 26.4 GB | 30.5 GB |
| UD-Q4_K_M | 22.1 GB | - | 23.1 GB | 23.3 GB | 23.5 GB | 24.0 GB | 25.0 GB | 27.1 GB | 31.2 GB |
| Q4_K_M Qwen_Qwen3.6-35B-A3B-Q4_K_M | 22.3 GB | Small | 23.3 GB | 23.4 GB | 23.7 GB | 24.2 GB | 25.2 GB | 27.2 GB | 31.3 GB |
| UD-Q4_K_XL Qwen3.6-35B-A3B-UD-Q4_K_XL | 22.4 GB | - | 23.4 GB | 23.5 GB | 23.7 GB | 24.3 GB | 25.3 GB | 27.3 GB | 31.4 GB |
| Q4_K_L | 22.7 GB | - | 23.7 GB | 23.8 GB | 24.0 GB | 24.6 GB | 25.6 GB | 27.6 GB | 31.7 GB |
| Q4_1 Qwen_Qwen3.6-35B-A3B-Q4_1 | 22.9 GB | Medium | 23.9 GB | 24.0 GB | 24.3 GB | 24.8 GB | 25.8 GB | 27.8 GB | 31.9 GB |
| BF16 Kimi-VL-A3B-Thinking-2506-bf16 | 31.9 GB | Original precision | 32.9 GB | 33.1 GB | 33.3 GB | 33.8 GB | 34.8 GB | 36.9 GB | 41.0 GB |
| BF16 Kimi-VL-A3B-Instruct-bf16 | 31.9 GB | Original precision | 32.9 GB | 33.1 GB | 33.3 GB | 33.8 GB | 34.8 GB | 36.9 GB | 41.0 GB |
| BF16 (2 parts) qwen3.8-27b/BF16/Qwen3.8-27B-BF16 | 54.7 GB | Original precision | 55.7 GB | 55.8 GB | 56.0 GB | 56.5 GB | 57.6 GB | 59.6 GB | 63.7 GB |
| BF16 (2 parts) Muse-Glimmer-30B-bf16 | 55.7 GB | Original precision | 56.7 GB | 56.9 GB | 57.1 GB | 57.6 GB | 58.6 GB | 60.7 GB | 64.8 GB |
| BF16 (2 parts) bf16 | 61.1 GB | Original precision | 62.1 GB | 62.2 GB | 62.5 GB | 63.0 GB | 64.0 GB | 66.0 GB | 70.1 GB |
| BF16 (2 parts) google_gemma-4-31B-it-bf16 | 61.4 GB | Original precision | 62.4 GB | 62.5 GB | 62.8 GB | 63.3 GB | 64.3 GB | 66.4 GB | 70.4 GB |
| BF16 (2 parts) Qwen_Qwen3.6-35B-A3B-bf16 | 71.1 GB | Original precision | 72.1 GB | 72.2 GB | 72.4 GB | 73.0 GB | 74.0 GB | 76.0 GB | 80.1 GB |
The mmproj file for image input (F16) is 0.64 GB Measured . It is not included in the memory above, so add it if you use images.
Ollama ollama run hf.co/elichen-skymizer/GAP-models:Q4_K_M
llama.cpp llama-server -hf elichen-skymizer/GAP-models:Q4_K_M
How to run
Mainline llama.cpp can load this model's format (gemma4) from b8637 (Apr 2, 2026) onward. Ollama and LM Studio bundle llama.cpp internally, so an older app fails to load it with errors such as "unknown model architecture". If that happens, update the app.
Example commands using Q4_K_M. To use another quantization, replace the name with one from the table above. How to run local LLMs
Ollama
Run it by pointing directly at the GGUF on Hugging Face.
ollama run hf.co/elichen-skymizer/GAP-models:Q4_K_M
llama.cpp
Pass the repository and quantization name to llama-server (or llama-cli) with -hf.
llama-server -hf elichen-skymizer/GAP-models:Q4_K_M
LM Studio
Paste elichen-skymizer/GAP-models into the search box, download it, and choose the quantization from the file list.
Models of a similar size (other developers)
-
27.3B
Qwen
Qwen3.8-27B
- UD-Q4_K_M size
- 16.5 GB
- Memory at 8K
- 18.2 GB
Measured Estimated VRAM 24 GB+
-
35.5B
huihui-ai
Huihui-Qwen3.6-35B-A3B-abliterated-MTP
- Q2_K size
- 13.2 GB
- Memory at 8K
- 14.5 GB
Measured Estimated VRAM 16 GB+
-
35.5B
ornith-ai
Ornith-1.5-35B-A3B
- Q4_K_M size
- 21.7 GB
- Memory at 8K
- 23.0 GB
Measured Estimated VRAM 24 GB+
-
27.3B
huihui-ai
Huihui-Qwen3.8-27B-abliterated
- Q4_K_M size
- 16.6 GB
- Memory at 8K
- 18.3 GB
Measured Estimated VRAM 24 GB+
-
25.2B
farbodtavakkoli
OTel-2.0-Decision-26B-A4B-4bit
- Q4_K_M size
- 16.8 GB
- Memory at 8K
- 18.4 GB
Measured Estimated VRAM 24 GB+
-
27.3B
JonathanColetti
Qwen3.8-27B-Uncensored
- Q4_K_M size
- 16.5 GB
- Memory at 8K
- 18.3 GB
Measured Estimated VRAM 24 GB+
Frequently asked questions
- How much memory do I need to run GAP-models (Q4_K_M)?
- The file size is 2.74 GB (measured). With a 8,192-token context, the estimated memory including the KV cache is 3.87 GB. The formula is published at https://localllm.az-arte.com/en/guide/memory-calculation.
- Does GAP-models run in 8 GB, 12 GB, 16 GB or 24 GB of VRAM?
- 8 GB: up to UD-Q4_K_XL (est. 7.31 GB); 12 GB: up to Q4_K_M (est. 11.7 GB); 16 GB: up to UD-Q4_K_XL (est. 17.0 GB); 24 GB: up to Q4_1 (est. 24.0 GB) (estimated at 8,192 tokens with an f16 KV cache, files from the most-downloaded source).
- What is the command to run GAP-models with Ollama?
- Run "ollama run hf.co/elichen-skymizer/GAP-models:Q4_K_M" to use the GGUF on Hugging Face directly. To use another quantization, replace the name at the end.