PR

Hy4 preview

PreviewGGUF by AngelSlimUpdated 2026-09-01
Download on Hugging Face
Parameters770B
Context1024K
Layers78
Architecturehyv4
Smallest quant229.4 GB
Quick answer

Hy4 preview needs about 469.2 GB of memory at Q4_K_M (a 467.3 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 3 quants, from 229.4 GB to 467.3 GB (Q4_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 5.6 GB.

Hy4 preview – a large hybrid‑MOE model for high‑capacity inference

Hy4 preview is a 770 billion‑parameter hybrid‑MOE model with a 1 048 576‑token context. It is distributed in three GGUF builds, each requiring a patched llama.cpp. The model supports English and Chinese.

Good at

  • Large context handling
  • Hybrid‑MOE architecture
  • Mixed‑precision flexibility

Limits

  • Requires patched llama.cpp
  • No stock llama.cpp support
  • Large VRAM needed (≈435 GiB for Q4_K_M)

Best for

  • Long‑form content generation
  • Complex reasoning tasks
  • Research prototyping

Who it’s for. Engineers and researchers who need a very large context model and have access to high‑end GPUs. The model is suitable for those willing to patch llama.cpp and manage large memory footprints.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—229.4 GB 231.3 GB
Too big Get file →
IQ1_M235.4 GB 237.2 GB
Too big Get file →
Q4_K_M467.3 GB 469.2 GB
Too big Get file →

For 8 GB VRAM, the STQ1_0 build is the only viable option, estimating ~214 GiB VRAM at full residency. For 16 GB VRAM, use the IQ1_M build (~237 GiB). The Q4_K_M build requires ~231 GiB and is best for maximum precision when memory allows.

What GPU do you need for Hy4 preview?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: no quant fits at an 8K context.

Which Hy4 preview quant should you choose?

Q4_K_M is the usual starting point for Hy4 preview: it balances file size and quality well. Here it is a 467.3 GB file.

If memory is tight, the smallest file is the first one in the table at 229.4 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturehyv4
Parameters770B (approximate, mixture of experts)
Layers78
Trained context1,048,576 tokens
KV cache (fp16)166 KB per token

Source

Quantized byAngelSlim
FamilyPreview
Downloads (HF)194,035
Last updated2026-09-01

Frequently asked questions about Hy4 preview

How much VRAM does Hy4 preview need?

About 469.2 GB at Q4_K_M with an 8K context window. That is the 467.3 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Hy4 preview on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 231.3 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Hy4 preview GGUF?

It depends on the quant: file is 229.4 GB, IQ1_M is 235.4 GB, Q4_K_M is 467.3 GB.

What context length does Hy4 preview support?

The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 1.4 GB at 8K and 5.6 GB at 32K.

Where can I download Hy4 preview GGUF?

The files are on Hugging Face in the AngelSlim/Hy4-preview-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Can I run Hy4 preview on 8 GB VRAM?

No. The smallest build still needs about 214 GiB VRAM at full residency.

Do I need a special llama.cpp build?

Yes, a patched llama.cpp with the hy4-preview-patch is required.

What languages does Hy4 preview support?

English and Chinese.

Similar models

Browse all →

More from AngelSlim

All by AngelSlim →

More Preview models

All Preview →

Other large models

All large models →