Filter by type:
Filter by tags:
hy4-preview
# Hy4-preview GGUF
Three GGUF builds of Hy4-Preview: https://huggingface.co/tencent/Hy4-preview
**Language / 语言:** English · 中文
**Neither file runs on stock llama.cpp.** The `hyv4` architecture is not upstream. Apply the
patches in `hy4-preview-patch/`
## English
### 1. What these are
**`Hy4-preview-Q4_K_M.gguf`** — a conventional Q4_K_M. Most tensors are Q4_K; `ffn_down_exps`
gets Q6_K on 37 layers via llama.cpp's own logic. Use this unless you are memory-constrained.
**`Hy4-preview-UD-IQ1_M.gguf`** - mixed precision with UD-IQ1_M strategy at ~2.44 bpw, roughly **half the size** for the
same model. The routed-expert `gate`/`up` projections run at 1.75 bpw (IQ1_M) and 2.0625 bpw (IQ2_XXS).
**`Hy4-preview-STQ1_0.gguf`** — mixed precision with MIX-STQ1_0 strategy at ~2.38 bpw, roughly **half the size** for the
same model. The routed-expert `gate`/`up` projections run at 1.3125 bpw (STQ1_0) on 29 layers and
2.0625 bpw (IQ2_XXS) on the other 48. See section 3.
### 2. Running them
Build a patched llama.cpp
```bash
git clone https://github.com/ggml-org/llama.cpp && cd llama.cpp
git checkout 0cea36222
...
Repository: localai
s1-mini-f16
S1-mini by Superwhisper in the publisher's 1.4 GB F16 GGUF format. This
variant preserves full model fidelity for hosts with enough memory.
Repository: localaiLicense: s1-mini-license
llm-jp-4-33b-thinking-bf16
LLM-jp-4-33B-thinking in the official 66.4 GB BF16 GGUF format. This
variant preserves the original model precision for hosts with enough
memory.
Repository: localaiLicense: apache-2.0
qwen3.8-flash-next-q8
Qwen3.8-Flash-Next in the higher-quality Q8_0 GGUF format, with the shared
BF16 vision projector. This build preserves more model quality but needs
more memory than the default Q4 variant.
Repository: localaiLicense: other
qwen3.8-flash-next-atomic-iq4
Qwen3.8 Flash Next in AtomicChat's AD-3.84bpw IQ4_XS M64 GGUF
build, with the F16 vision projector. The n-gram table occupies a
separate shard. This entry enables memory mapping and disables llama.cpp
automatic parameter fitting as required by the publisher.
Repository: localaiLicense: other
qwen3.8-flash-next-atomic-q4
Qwen3.8 Flash Next in AtomicChat's AD-4.27bpw Q4_K_M M64 GGUF
build, with the F16 vision projector. The n-gram table occupies a
separate shard. This entry enables memory mapping and disables llama.cpp
automatic parameter fitting as required by the publisher.
Repository: localaiLicense: other
ling-3.0-tiny-q8
Ling-3.0-tiny in the higher-quality Q8_0 GGUF format. This variant
preserves more model fidelity for hosts with enough memory.
Repository: localaiLicense: mit
hy-mt2-1.8b-q8
Hy-MT2-1.8B in the higher-quality 1.9 GB Q8_0 GGUF format. This variant
preserves more model fidelity for hosts with enough memory.
Repository: localaiLicense: apache-2.0
ling-3.0-flash-iq2
Ling-3.0-flash in the higher-quality 44.7 GB AD-IQ2_XS GGUF format. This
variant preserves more model fidelity for hosts with enough memory.
Repository: localaiLicense: mit
qwen3.8-27b-q8
Qwen3.8-27B in the official Q8_0 GGUF format. This variant provides higher
model fidelity for hosts with enough memory.
Repository: localaiLicense: apache-2.0
underdog-saluki-27b-1.0
Underdog Saluki 27B 1.0 is a mixed IQ2 quantization of Qwen3.8-27B
for chat, reasoning, and tool use. This llama.cpp build includes the
F16 vision projector and uses the embedded chat template.
The weights and projector total approximately 8.83 GB before runtime memory.
Repository: localaiLicense: apache-2.0
qwen3.8-9b-q8
Qwen3.8-9B in the higher-quality Q8_0 GGUF format. This variant preserves
more model fidelity for hosts with enough memory.
Repository: localaiLicense: apache-2.0
qwen3.8-4b-q8
Qwen3.8-4B in the higher-quality Q8_0 GGUF format. This variant preserves
more model fidelity for hosts with enough memory.
Repository: localaiLicense: apache-2.0
twil-lm3-q8
TwIL-LM3 in the publisher's near-lossless Q8_0 GGUF format for
quality-sensitive use on hosts with enough memory.
Repository: localaiLicense: webai-non-commercial-license-ver.-1.0
nemotron-3.5-lightning-30b-a3b-q8
NVIDIA Nemotron 3.5 Lightning 30B-A3B in the official high-quality Q8_0
GGUF format for hosts with enough memory.
Repository: localaiLicense: openmdw-1.1
muse-glimmer-30b-17gb
Muse Glimmer's smaller 17 GB K-quant GGUF with the official quantized
perception encoder. It preserves the model's agentic, coding, tool-use,
multilingual, and image-understanding capabilities for hosts with less
memory than the dynamic quantization requires.
Repository: localaiLicense: apache-2.0
muse-glimmer-30b-17gb-dflash
Muse Glimmer's smaller 17 GB K-quant GGUF with the official quantized
perception encoder and DFlash drafter. This is the lowest-memory published
build that retains image understanding and block-speculative decoding.
Flash attention is enabled for the DFlash path.
Repository: localaiLicense: apache-2.0
b1-9b-q8
B1-9B in Q8_0 GGUF format with the Q8_0 vision projector. This variant
preserves more weight precision and requires more memory than Q4_K_M.
Repository: localaiLicense: apache-2.0
qwen3.6-35b-a3b-genesis-hermes-v7
Qwen3.6-35B-A3B Genesis Hermes V7 is LuffyTheFox's Apache-2.0 multimodal,
agentic derivative of HauhauCS's uncensored Qwen3.6-35B-A3B model. It
combines Genesis tensor calibration with Hermes function-calling data while
retaining the 35B mixture-of-experts architecture, roughly 3B active
parameters per token, and the native 262K-token context window.
This entry's own payload uses the model card's recommended APEX GGUF and the
shared F16 multimodal projector. Automatic variant selection may instead
choose Compact APEX, an MTP-enabled APEX build, or Q8_K_P based on serving
features and available memory. The model card recommends Jinja chat
templates and at least a 128K context for its thinking behavior.
Repository: localaiLicense: apache-2.0
qwen3.6-35b-a3b-genesis-hermes-v7-apex-compact
Qwen3.6-35B-A3B Genesis Hermes V7 in the smaller APEX Compact GGUF format,
with the shared F16 multimodal projector. This build preserves the model's
multimodal, reasoning, coding, and agentic capabilities for hosts with less
memory than the recommended full APEX build.
Repository: localaiLicense: apache-2.0
laguna-xs-2.1-apex-i-balanced
Laguna XS 2.1 in the 24.3 GB APEX-I Balanced format, an importance-matrix build balancing fidelity and memory use for llama.cpp.
License: OpenMDW 1.1.
Repository: localaiLicense: other
Page 1