TE

Tencent Hy 30B A3B Uncensored Heretic

TencentGGUF by mradermacherUpdated 2026-09-01
Download on Hugging Face
Parameters30.1B
Context256K
Layers48
Architecturehy_v3
Smallest quant0.1 GB
Quick answer

Tencent Hy 30B A3B Uncensored Heretic needs about 19.5 GB of memory at Q4_K_M (a 18.2 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 24 quants, from 0.1 GB to 24.7 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 3.2 GB.

Tencent Hy 30B A3B Uncensored Heretic – 30.1B parameter MoE model

Tencent Hy 30B A3B Uncensored Heretic is a 30.1B parameter mixture of experts model based on hy_v3 architecture. It supports 262144 token context and is released under apache-2.0. The model is intended for research and safety studies.

Good at

  • reasoning
  • tools usage
  • multilingual translation

Limits

  • no vision support
  • requires large VRAM for full precision
  • unverified safety alignment

Best for

  • research on MoE models
  • safety and alignment testing
  • red‑team evaluation

Who it’s for. Researchers and developers who need a large multilingual MoE model for experimentation. The model is not recommended for public deployment due to reduced safety alignment.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.1 GB 1.4 GB
Fits Get file →
IQ1_S6.2 GB 7.5 GB
Tight Get file →
IQ1_M6.9 GB 8.2 GB
Too big Get file →
IQ2_XXS8.0 GB 9.3 GB
Too big Get file →
IQ2_XS8.9 GB 10.2 GB
Too big Get file →
IQ2_S9.0 GB 10.3 GB
Too big Get file →
IQ2_M9.9 GB 11.2 GB
Too big Get file →
Q2_K_S10.3 GB 11.6 GB
Too big Get file →
Q2_K11.1 GB 12.4 GB
Too big Get file →
IQ3_XXS11.6 GB 13.0 GB
Too big Get file →
IQ3_XS12.4 GB 13.7 GB
Too big Get file →
Q3_K_S13.1 GB 14.4 GB
Too big Get file →
IQ3_S13.1 GB 14.4 GB
Too big Get file →
IQ3_M13.3 GB 14.6 GB
Too big Get file →
Q3_K_M14.4 GB 15.8 GB
Too big Get file →
Q3_K_L15.6 GB 16.9 GB
Too big Get file →
IQ4_XS16.1 GB 17.4 GB
Too big Get file →
Q4_017.1 GB 18.4 GB
Too big Get file →
Q4_K_S17.2 GB 18.5 GB
Too big Get file →
Q4_K_M18.2 GB 19.5 GB
Too big Get file →
Q4_118.9 GB 20.2 GB
Too big Get file →
Q5_K_S20.7 GB 22.1 GB
Too big Get file →
Q5_K_M21.4 GB 22.7 GB
Too big Get file →
Q6_K24.7 GB 26.0 GB
Too big Get file →

For 8 GB VRAM, use Q4_K_M (18.24 GB file, 19.5 GB memory at 8k). For 12 GB VRAM, Q4_K (17.09 GB file, 18.4 GB memory). For 16 GB VRAM, Q5_K_M (21.37 GB file, 22.7 GB memory). Estimates are approximate.

What GPU do you need for Tencent Hy 30B A3B Uncensored Heretic?

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: the best quant that fits is the main file (0.1 GB file, about 1.4 GB in use).
  • 6 GB VRAM: the best quant that fits is the main file (0.1 GB file, about 1.4 GB in use).
  • 8 GB VRAM: the best quant that fits is IQ1_S (6.2 GB file, about 7.5 GB in use).
  • 12 GB VRAM: the best quant that fits is Q2_K_S (10.3 GB file, about 11.6 GB in use).
  • 16 GB VRAM: the best quant that fits is Q3_K_M (14.4 GB file, about 15.8 GB in use).
  • 24 GB VRAM: the best quant that fits is Q5_K_M (21.4 GB file, about 22.7 GB in use).
  • 32 GB VRAM: the best quant that fits is Q6_K (24.7 GB file, about 26.0 GB in use).

Which Tencent Hy 30B A3B Uncensored Heretic quant should you choose?

Q4_K_M is the usual starting point for Tencent Hy 30B A3B Uncensored Heretic: it balances file size and quality well. Here it is a 18.2 GB file.

If memory is tight, the smallest file is the first one in the table at 0.1 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

Architecturehy_v3
Parameters30.1B (approximate, mixture of experts)
Active parameters3B
Layers48
Trained context262,144 tokens
KV cache (fp16)96 KB per token

Source

Quantized bymradermacher
FamilyTencent
Licenseapache-2.0
Downloads (HF)959
Last updated2026-09-01

Frequently asked questions about Tencent Hy 30B A3B Uncensored Heretic

How much VRAM does Tencent Hy 30B A3B Uncensored Heretic need?

About 19.5 GB at Q4_K_M with an 8K context window. That is the 18.2 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 Tencent Hy 30B A3B Uncensored Heretic on an 8 GB GPU?

Yes, with IQ1_S: the 6.2 GB file needs about 7.5 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Tencent Hy 30B A3B Uncensored Heretic GGUF?

It depends on the quant: file is 0.1 GB, IQ1_S is 6.2 GB, IQ1_M is 6.9 GB, IQ2_XXS is 8.0 GB, IQ2_XS is 8.9 GB, IQ2_S is 9.0 GB and 18 more.

What context length does Tencent Hy 30B A3B Uncensored Heretic support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.2 GB at 32K.

Where can I download Tencent Hy 30B A3B Uncensored Heretic GGUF?

The files are on Hugging Face in the mradermacher/Tencent-Hy-30B-A3B-uncensored-heretic-i1-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Tencent Hy 30B A3B Uncensored Heretic use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run this model on 8 GB VRAM?

Use the Q4_K_M quant, which requires about 19.5 GB memory at 8k context. 8 GB VRAM is insufficient for full inference.

Is this model safe for public use?

The model has reduced safety alignment and is intended only for research and experimentation. Do not deploy it in public services.

What license applies to this model?

The model is released under the apache-2.0 license.

Other GGUFs of Tencent-Hy-30B-A3B-uncensored-heretic

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Tencent models

All Tencent →

Other medium models

All medium models →