SW

Swish 26B A4B

SwishGGUF by ReadyArtUpdated 2026-10-19
Download on Hugging Face
Parameters25.2B
Context256K
Layers30
Architecturegemma4
Smallest quant0.1 GB
Quick answer

Swish 26B A4B needs about 26.2 GB of memory at Q4_K_M (a 17.7 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 12 quants, from 0.1 GB to 27.6 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 32.2 GB.

Swish 26B A4B – a 26 billion‑parameter gemma4 model with Mixture of Experts

Swish 26B A4B is a 26 billion‑parameter model built on the gemma4 architecture. It includes 4 billion active parameters in a mixture of experts. The model was trained on 262,144 context tokens and is released under the Apache‑2.0 license.

Good at

  • Roleplay capabilities
  • Large context handling
  • Mixture of experts support

Limits

  • No vision support
  • Requires large VRAM for full precision
  • Limited language coverage

Best for

  • Interactive storytelling
  • Complex dialogue generation
  • Role‑playing scenarios

Who it’s for. Developers and researchers needing a large, context‑rich language model with roleplay abilities. The model is suitable for applications that can accommodate its memory requirements and do not require vision or extensive language coverage.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.1 GB 8.7 GB
Too big Get file →
IQ2_XS10.8 GB 19.3 GB
Too big Get file →
IQ3_XS12.5 GB 21.1 GB
Too big Get file →
Q3_K_M14.2 GB 22.7 GB
Too big Get file →
IQ4_XS14.8 GB 23.3 GB
Too big Get file →
IQ4_NL15.3 GB 23.9 GB
Too big Get file →
Q4_K_S16.3 GB 24.9 GB
Too big Get file →
Q4_K_M17.7 GB 26.2 GB
Too big Get file →
Q5_K_S18.9 GB 27.4 GB
Too big Get file →
Q5_K_M20.0 GB 28.6 GB
Too big Get file →
Q6_K23.5 GB 32.1 GB
Too big Get file →
Q8_027.6 GB 36.1 GB
Too big Get file →

For 8 GB VRAM, use Q4_K_M (17.67 GB file, ~26.2 GB memory). For 12 GB VRAM, Q5_K_S (18.86 GB file, ~27.4 GB memory). For 16 GB VRAM, Q6_K (23.51 GB file, ~32.1 GB memory). Estimates are approximate.

What GPU do you need for Swish 26B A4B?

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 8.7 GB in use).
  • 16 GB VRAM: the best quant that fits is the main file (0.1 GB file, about 8.7 GB in use).
  • 24 GB VRAM: the best quant that fits is IQ4_NL (15.3 GB file, about 23.9 GB in use).
  • 32 GB VRAM: the best quant that fits is Q5_K_M (20.0 GB file, about 28.6 GB in use).
  • 48 GB VRAM: the best quant that fits is Q8_0 (27.6 GB file, about 36.1 GB in use).

Which Swish 26B A4B quant should you choose?

Q4_K_M is the usual starting point for Swish 26B A4B: it balances file size and quality well. Here it is a 17.7 GB file.

The Q8_0 file is 27.6 GB, about 1.6 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

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 byReadyArt
FamilySwish
Licenseapache-2.0
Downloads (HF)4,602
Last updated2026-10-19

Frequently asked questions about Swish 26B A4B

How much VRAM does Swish 26B A4B need?

About 26.2 GB at Q4_K_M with an 8K context window. That is the 17.7 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 Swish 26B A4B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 8.7 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Swish 26B A4B GGUF?

It depends on the quant: file is 0.1 GB, IQ2_XS is 10.8 GB, IQ3_XS is 12.5 GB, Q3_K_M is 14.2 GB, IQ4_XS is 14.8 GB, IQ4_NL is 15.3 GB and 6 more.

What context length does Swish 26B A4B 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 Swish 26B A4B GGUF?

The files are on Hugging Face in the ReadyArt/Swish-26B-A4B-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 Swish 26B A4B 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 Swish 26B A4B on 8 GB VRAM?

Use the Q4_K_M quantized file; it requires about 26.2 GB of VRAM at 8 k context, so 8 GB is insufficient. A larger VRAM or lower context is needed.

Does Swish 26B A4B support image inputs?

No, the model does not have vision capabilities.

What license applies to Swish 26B A4B?

The model is released under the Apache‑2.0 license.

Other GGUFs of Swish-26B-A4B

Similar models

Browse all →

More from ReadyArt

All by ReadyArt →

More Swish models

All Swish →

Other medium models

All medium models →