Skip to content
localllm
日本語

ukisai

Swift-1.5-Qwen3.8-Flash-Next

Image input

Parameters
176.9B
Max context length
262,144
Architecture
qwen4exp
License
Conditional
Base model
ukisai/Swift-Qwen3.8-Flash-Next
License
swift-open-license-1.0
llama.cpp
b10660+

About Swift-1.5-Qwen3.8-Flash-Next

Swift-1.5-Qwen3.8-Flash-Next is a GGUF model based on ukisai/Swift-Qwen3.8-Flash-Next on Hugging Face. GGUF files are published by ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF. It has about 176.9B parameters, the qwen4exp architecture, a maximum context length of 262,144 tokens (values written in the GGUF).

There are 27 quantizations, with file sizes from 70.1 GB (IQ1_S) to 188 GB (Q8_0) (actual sizes on Hugging Face).

Running Q4_K_M (120 GB) with an 8,192-token context needs an estimated 121 GB of memory, and an estimated 122 GB at 32,768 tokens. For 8,192 tokens, a GPU with 128 GB or more of VRAM, or a Mac with 128 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, IQ1_S (70.1 GB), needs an estimated 71.5 GB at 8,192 tokens.

The original model is licensed under Swift Open License v1.0, which allows commercial use with conditions such as scale or purpose. Check the original license for the terms of use. An mmproj file for image input is also available.

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

GGUF source

ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF

Published by the original developer About this label

Last checked: Nov 11, 2026

Architecture
qwen4exp
llama.cpp b10660 and later
Parameters
176.9B
Max context length
262,144
License
swift-open-license-1.0 Conditional
File size and memory per quantization for Swift-1.5-Qwen3.8-Flash-Next (ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF)
Quantization File size Measured Quality guide 4,096 8,192 16,384 32,768 65,536 131,072 262,144 max
IQ1_S (2 parts) 70.1 GB Very large 71.4 GB 71.5 GB 71.7 GB 72.2 GB 73.1 GB 74.9 GB 78.5 GB
IQ1_M (2 parts) 72.0 GB Very large 73.3 GB 73.4 GB 73.6 GB 74.1 GB 75.0 GB 76.8 GB 80.4 GB
IQ2_XXS (2 parts) 75.2 GB Very large 76.5 GB 76.6 GB 76.8 GB 77.3 GB 78.2 GB 80.0 GB 83.6 GB
IQ2_XS (2 parts) 77.7 GB Very large 79.0 GB 79.1 GB 79.4 GB 79.8 GB 80.7 GB 82.5 GB 86.1 GB
IQ2_S (2 parts) 77.8 GB Very large 79.1 GB 79.2 GB 79.4 GB 79.9 GB 80.8 GB 82.6 GB 86.2 GB
IQ2_M (2 parts) 80.4 GB Large 81.6 GB 81.8 GB 82.0 GB 82.4 GB 83.3 GB 85.2 GB 88.8 GB
Q2_K (2 parts) 80.9 GB Very large 82.2 GB 82.3 GB 82.5 GB 83.0 GB 83.9 GB 85.7 GB 89.3 GB
IQ3_XXS (2 parts) 88.0 GB Large 89.3 GB 89.4 GB 89.6 GB 90.1 GB 91.0 GB 92.8 GB 96.4 GB
Q3_K_S (2 parts) 89.4 GB Large 90.7 GB 90.8 GB 91.1 GB 91.5 GB 92.4 GB 94.2 GB 97.9 GB
IQ3_XS (3 parts) 91.9 GB Large 93.2 GB 93.3 GB 93.6 GB 94.0 GB 94.9 GB 96.7 GB 100 GB
Q3_K_M (3 parts) 92.0 GB Medium 93.2 GB 93.3 GB 93.6 GB 94.0 GB 94.9 GB 96.7 GB 100 GB
Q3_K_L (3 parts) 93.2 GB Medium 94.5 GB 94.6 GB 94.8 GB 95.3 GB 96.2 GB 98.0 GB 102 GB
IQ3_M (3 parts) 93.2 GB Medium 94.5 GB 94.6 GB 94.8 GB 95.3 GB 96.2 GB 98.0 GB 102 GB
IQ4_XS (3 parts) 97.7 GB Small 99.0 GB 99.1 GB 99.3 GB 99.8 GB 101 GB 102 GB 106 GB
IQ4_NL (3 parts) 100 GB Small 102 GB 102 GB 102 GB 102 GB 103 GB 105 GB 109 GB
Q4_0 (3 parts) 101 GB Medium 102 GB 102 GB 102 GB 103 GB 104 GB 105 GB 109 GB
Q2_K_L (4 parts) 107 GB - 108 GB 109 GB 109 GB 109 GB 110 GB 112 GB 116 GB
Q4_1 (3 parts) 111 GB Medium 113 GB 113 GB 113 GB 113 GB 114 GB 116 GB 120 GB
Q4_K_S (3 parts) 113 GB Medium 114 GB 115 GB 115 GB 115 GB 116 GB 118 GB 122 GB
Q3_K_XL (4 parts) 119 GB - 121 GB 121 GB 121 GB 121 GB 122 GB 124 GB 128 GB
Q4_K_M (3 parts) 120 GB Small 121 GB 121 GB 121 GB 122 GB 123 GB 124 GB 128 GB
Q5_K_S (3 parts) 128 GB Small 129 GB 129 GB 130 GB 130 GB 131 GB 133 GB 137 GB
Q5_K_M (4 parts) 135 GB Very small 136 GB 136 GB 136 GB 137 GB 138 GB 139 GB 143 GB
Q4_K_L (4 parts) 139 GB - 141 GB 141 GB 141 GB 141 GB 142 GB 144 GB 148 GB
Q5_K_L (5 parts) 151 GB - 152 GB 152 GB 153 GB 153 GB 154 GB 156 GB 159 GB
Q6_K (5 parts) 168 GB Very small 169 GB 170 GB 170 GB 170 GB 171 GB 173 GB 177 GB
Q8_0 (5 parts) 188 GB Nearly identical 190 GB 190 GB 190 GB 190 GB 191 GB 193 GB 197 GB

The mmproj file for image input (BF16) is 0.91 GB Measured . It is not included in the memory above, so add it if you use images. 

Ollama  ollama run hf.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
llama.cpp  llama-server -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

License and commercial use

Swift Open License v1.0 Conditional

Commercial use by organizations with total revenue of USD 1 million or more in the most recent fiscal year is not permitted under this license (a separate agreement is needed). Below that, commercial use is allowed.

Based on the license on the original model's Hugging Face page. Original license text (checked Oct 18, 2026)

The summary states facts this site read from the original text and is not legal advice. Always check the original license and its latest terms before use. All licenses

How to run

Mainline llama.cpp can load this model's format (qwen4exp) from b10660 (Aug 27, 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/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

llama.cpp

Pass the repository and quantization name to llama-server (or llama-cli) with -hf.

llama-server -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

LM Studio

Paste ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF into the search box, download it, and choose the quantization from the file list.

Frequently asked questions

How much memory do I need to run Swift-1.5-Qwen3.8-Flash-Next (Q4_K_M)?
The file size is 120 GB (measured). With a 8,192-token context, the estimated memory including the KV cache is 121 GB. The formula is published at https://localllm.az-arte.com/en/guide/memory-calculation.
Can Swift-1.5-Qwen3.8-Flash-Next be used commercially?
The original model's license is Swift Open License v1.0 (Commercial use with conditions).Commercial use by organizations with total revenue of USD 1 million or more in the most recent fiscal year is not permitted under this license (a separate agreement is needed). Below that, commercial use is allowed. Check the original license text before deciding.
Does Swift-1.5-Qwen3.8-Flash-Next run in 8 GB, 12 GB, 16 GB or 24 GB of VRAM?
8 GB: no quantization fits; 12 GB: no quantization fits; 16 GB: no quantization fits; 24 GB: no quantization fits (estimated at 8,192 tokens with an f16 KV cache, files from the most-downloaded source).
What is the command to run Swift-1.5-Qwen3.8-Flash-Next with Ollama?
Run "ollama run hf.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M" to use the GGUF on Hugging Face directly. To use another quantization, replace the name at the end.