ZE

BA Zephyria 39b i1

ZephyriaGGUF by mradermacherUpdated 2025-03-03
Download on Hugging Face
Parameters39B
Context128K
Layers99
Architecturellama
Smallest quant8.4 GB
Quick answer

BA Zephyria 39b i1 needs about 27.2 GB of memory at Q4_K_M (a 23.4 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 21 quants, from 8.4 GB (IQ1_S) to 32.0 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 13.3 GB.

BA Zephyria 39b i1 – a 39‑billion‑parameter Llama model in GGUF format

BA Zephyria 39b i1 is a 39‑billion‑parameter Llama model with a 131072‑token context window. It is available in multiple quantizations ranging from IQ1_S to Q6_K. The model is derived from SteelStorage/BA‑Zephyria‑39b and is distributed via the Hugging Face GGUF repository.

Good at

  • Large‑context reasoning
  • High‑capacity text generation
  • Fine‑grained language tasks

Limits

  • No vision support
  • License not specified
  • Requires 8‑k context for memory estimates

Best for

  • Research prototypes
  • Long‑form content creation
  • Complex dialogue systems

Who it’s for. Researchers and developers needing a large‑scale Llama model with flexible quantization options. The model is suitable for experimentation and deployment on systems with sufficient VRAM and RAM.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S8.4 GB 12.2 GB
Too big Get file →
IQ1_M9.1 GB 13.0 GB
Too big Get file →
IQ2_XXS10.4 GB 14.3 GB
Too big Get file →
IQ2_XS11.6 GB 15.4 GB
Too big Get file →
IQ2_S12.3 GB 16.1 GB
Too big Get file →
IQ2_M13.3 GB 17.1 GB
Too big Get file →
Q2_K14.4 GB 18.3 GB
Too big Get file →
IQ3_XXS15.0 GB 18.9 GB
Too big Get file →
IQ3_XS16.0 GB 19.9 GB
Too big Get file →
Q3_K_S16.9 GB 20.7 GB
Too big Get file →
IQ3_S16.9 GB 20.8 GB
Too big Get file →
IQ3_M17.6 GB 21.4 GB
Too big Get file →
Q3_K_M18.8 GB 22.7 GB
Too big Get file →
Q3_K_L20.5 GB 24.4 GB
Too big Get file →
IQ4_XS20.9 GB 24.7 GB
Too big Get file →
Q4_022.1 GB 25.9 GB
Too big Get file →
Q4_K_S22.2 GB 26.0 GB
Too big Get file →
Q4_K_M23.4 GB 27.2 GB
Too big Get file →
Q5_K_S26.9 GB 30.7 GB
Too big Get file →
Q5_K_M27.6 GB 31.4 GB
Too big Get file →
Q6_K32.0 GB 35.8 GB
Too big Get file →

For 8‑GB VRAM, use IQ1_S (file size 8.38 GB, estimated 12.2 GB memory). For 12‑GB VRAM, IQ2_S (12.26 GB, 16.1 GB memory) is a good balance. If 16‑GB VRAM is available, Q4_K_S (22.16 GB, 26 GB memory) offers higher quality with moderate memory use. Estimates are approximate.

What GPU do you need for BA Zephyria 39b i1?

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: the best quant that fits is IQ2_XS (11.6 GB file, about 15.4 GB in use).
  • 24 GB VRAM: the best quant that fits is Q3_K_M (18.8 GB file, about 22.7 GB in use).
  • 32 GB VRAM: the best quant that fits is Q5_K_M (27.6 GB file, about 31.4 GB in use).
  • 48 GB VRAM: the best quant that fits is Q6_K (32.0 GB file, about 35.8 GB in use).

Which BA Zephyria 39b i1 quant should you choose?

Q4_K_M is the usual starting point for BA Zephyria 39b i1: it balances file size and quality well. Here it is a 23.4 GB file.

If memory is tight, the smallest file is IQ1_S at 8.4 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters39B
Layers99
Trained context131,072 tokens
KV cache (fp16)396 KB per token

Source

Quantized bymradermacher
FamilyZephyria
Downloads (HF)1,617
Last updated2025-03-03

Frequently asked questions about BA Zephyria 39b i1

How much VRAM does BA Zephyria 39b i1 need?

About 27.2 GB at Q4_K_M with an 8K context window. That is the 23.4 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 BA Zephyria 39b i1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 12.2 GB, so you would need a larger GPU or CPU offloading.

What is the file size of BA Zephyria 39b i1 GGUF?

It depends on the quant: IQ1_S is 8.4 GB, IQ1_M is 9.1 GB, IQ2_XXS is 10.4 GB, IQ2_XS is 11.6 GB, IQ2_S is 12.3 GB, IQ2_M is 13.3 GB and 15 more.

What context length does BA Zephyria 39b i1 support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 3.3 GB at 8K and 13.3 GB at 32K.

Where can I download BA Zephyria 39b i1 GGUF?

The files are on Hugging Face in the mradermacher/BA-Zephyria-39b-i1-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 BA Zephyria 39b i1 on 8 GB VRAM?

Use the IQ1_S quant, which is 8.38 GB in size and requires about 12.2 GB of memory at 8‑k context.

What quant gives the best quality for a 16 GB VRAM system?

Q4_K_S is 22.16 GB and needs about 26 GB of memory, fitting within a 16‑GB VRAM setup when using swap or off‑load.

Is there a license restriction for using this model?

The license field is null; check the original repository for any usage terms.

Other GGUFs of BA-Zephyria-39b

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Zephyria models

All Zephyria →

Other large models

All large models →