GE

Gemma 4 26B A4B Heretic Stable

GemmaGGUF by mradermacherUpdated 2026-05-23
Download on Hugging Face
Parameters25.2B
Context256K
Layers30
Architecturegemma4
Smallest quant0.1 GB
Quick answer

Gemma 4 26B A4B Heretic Stable needs about 25.3 GB of memory at Q4_K_M (a 16.8 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 24 quants, from 0.1 GB to 22.6 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 32.2 GB.

Gemma 4 26B A4B Heretic Stable – 26B parameter Mixture‑of‑Experts model

Gemma 4 26B A4B Heretic Stable is a 26B parameter language model with a Mixture‑of‑Experts architecture. It supports 262,144 context tokens and is optimized for recent Transformers versions. The model is released under the Apache‑2.0 license.

Good at

  • Reasoning
  • Large context handling
  • Mixture‑of‑Experts inference

Limits

  • Requires high VRAM for full precision
  • No vision support in current GGUF
  • Large file size for higher‑quality quants

Best for

  • Research on large‑scale transformer behavior
  • High‑performance local inference
  • Prototyping Mixture‑of‑Experts setups

Who it’s for. Researchers and developers needing a large, reasoning‑capable model for local deployment. The model is suitable for advanced inference environments and is not intended for production use without further safety measures.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.1 GB 8.6 GB
Too big Get file →
IQ1_S8.3 GB 16.8 GB
Too big Get file →
IQ1_M8.7 GB 17.2 GB
Too big Get file →
IQ2_XXS9.3 GB 17.9 GB
Too big Get file →
IQ2_XS9.8 GB 18.4 GB
Too big Get file →
IQ2_S9.9 GB 18.4 GB
Too big Get file →
IQ2_M10.4 GB 18.9 GB
Too big Get file →
Q2_K10.6 GB 19.1 GB
Too big Get file →
Q2_K_S10.6 GB 19.2 GB
Too big Get file →
IQ3_XXS11.3 GB 19.9 GB
Too big Get file →
IQ3_XS11.6 GB 20.2 GB
Too big Get file →
IQ3_S12.2 GB 20.8 GB
Too big Get file →
Q3_K_S12.2 GB 20.8 GB
Too big Get file →
IQ3_M12.4 GB 20.9 GB
Too big Get file →
Q3_K_M13.3 GB 21.8 GB
Too big Get file →
Q3_K_L13.8 GB 22.4 GB
Too big Get file →
IQ4_XS13.9 GB 22.5 GB
Too big Get file →
Q4_014.5 GB 23.0 GB
Too big Get file →
Q4_K_S15.5 GB 24.0 GB
Too big Get file →
Q4_116.0 GB 24.5 GB
Too big Get file →
Q4_K_M16.8 GB 25.3 GB
Too big Get file →
Q5_K_S18.0 GB 26.5 GB
Too big Get file →
Q5_K_M19.1 GB 27.7 GB
Too big Get file →
Q6_K22.6 GB 31.2 GB
Too big Get file →

For 8 GB VRAM, use IQ1_S (8.29 GB file, 16.8 GB memory at 8 k context). For 12 GB VRAM, IQ2_S (9.87 GB file, 18.4 GB memory) or Q2_K (10.58 GB file, 19.1 GB memory). For 16 GB VRAM, Q4_K_M (16.8 GB file, 25.3 GB memory) offers a balance between size and quality.

What GPU do you need for Gemma 4 26B A4B Heretic Stable?

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: the best quant that fits is the main file (0.1 GB file, about 8.6 GB in use).
  • 16 GB VRAM: the best quant that fits is the main file (0.1 GB file, about 8.6 GB in use).
  • 24 GB VRAM: the best quant that fits is Q4_K_S (15.5 GB file, about 24.0 GB in use).
  • 32 GB VRAM: the best quant that fits is Q6_K (22.6 GB file, about 31.2 GB in use).

Which Gemma 4 26B A4B Heretic Stable quant should you choose?

Q4_K_M is the usual starting point for Gemma 4 26B A4B Heretic Stable: it balances file size and quality well. Here it is a 16.8 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

Architecturegemma4
Parameters25.2B (approximate, mixture of experts)
Active parameters4B
Layers30
Trained context262,144 tokens
KV cache (fp16)960 KB per token

Source

Quantized bymradermacher
FamilyGemma
Licenseapache-2.0
Downloads (HF)2,350
Last updated2026-05-23

Frequently asked questions about Gemma 4 26B A4B Heretic Stable

How much VRAM does Gemma 4 26B A4B Heretic Stable need?

About 25.3 GB at Q4_K_M with an 8K context window. That is the 16.8 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 Gemma 4 26B A4B Heretic Stable on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 8.6 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Gemma 4 26B A4B Heretic Stable GGUF?

It depends on the quant: file is 0.1 GB, IQ1_S is 8.3 GB, IQ1_M is 8.7 GB, IQ2_XXS is 9.3 GB, IQ2_XS is 9.8 GB, IQ2_S is 9.9 GB and 18 more.

What context length does Gemma 4 26B A4B Heretic Stable support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 8.1 GB at 8K and 32.2 GB at 32K.

Where can I download Gemma 4 26B A4B Heretic Stable GGUF?

The files are on Hugging Face in the mradermacher/gemma-4-26B-A4B-Heretic-Stable-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 Gemma 4 26B A4B Heretic Stable 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 IQ1_S quant, which requires 16.8 GB memory at 8 k context and a 8.29 GB file size.

Does this model support vision tasks?

No, the current GGUF files do not include vision capabilities.

What license applies to this model?

The model is released under the Apache‑2.0 license.

Other GGUFs of gemma-4-26B-A4B-Heretic-Stable

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Gemma models

All Gemma →

Other medium models

All medium models →