ER

ERNIE 4.5 21B A3B PT

ErnieGGUF by mradermacherUpdated 2025-07-18
Download on Hugging Face
Parameters21.8B
Context128K
Layers28
Architectureernie4_5-moe
Smallest quant4.5 GB
Quick answer

ERNIE 4.5 21B A3B PT needs about 14.2 GB of memory at Q4_K_M (a 13.2 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 23 quants, from 4.5 GB (IQ1_S) to 17.9 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 1.9 GB.

ERNIE 4.5 21B A3B PT – 21.8B parameter MoE LLM

ERNIE 4.5 21B A3B PT is a text‑only Mixture‑of‑Experts model with 21.8B total parameters and 3B active parameters. It supports 131072 token context and is released under Apache‑2.0. The model is available in multiple GGUF quantizations.

Good at

  • Large context handling
  • MoE efficient inference
  • Text generation

Limits

  • No vision support
  • Requires high VRAM for larger quants
  • Apache‑2.0 license

Best for

  • Long‑form content creation
  • Complex Q&A with extended context
  • Research prototyping

Who it’s for. Developers and researchers needing a high‑capacity, text‑focused MoE model. Suitable for environments with sufficient GPU memory and a need for extended context length.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S4.5 GB 5.4 GB
Fits Get file →
IQ1_M5.0 GB 5.9 GB
Fits Get file →
IQ2_XXS5.8 GB 6.8 GB
Fits Get file →
IQ2_XS6.5 GB 7.4 GB
Tight Get file →
IQ2_S6.5 GB 7.5 GB
Tight Get file →
IQ2_M7.2 GB 8.1 GB
Too big Get file →
Q2_K_S7.4 GB 8.4 GB
Too big Get file →
Q2_K8.1 GB 9.0 GB
Too big Get file →
IQ3_XXS8.5 GB 9.4 GB
Too big Get file →
IQ3_XS9.0 GB 10.0 GB
Too big Get file →
Q3_K_S9.5 GB 10.5 GB
Too big Get file →
IQ3_S9.5 GB 10.5 GB
Too big Get file →
IQ3_M9.6 GB 10.6 GB
Too big Get file →
Q3_K_M10.5 GB 11.4 GB
Too big Get file →
Q3_K_L11.4 GB 12.3 GB
Too big Get file →
IQ4_XS11.7 GB 12.7 GB
Too big Get file →
Q4_012.4 GB 13.4 GB
Too big Get file →
Q4_K_S12.4 GB 13.4 GB
Too big Get file →
Q4_K_M13.2 GB 14.2 GB
Too big Get file →
Q4_113.7 GB 14.7 GB
Too big Get file →
Q5_K_S15.1 GB 16.0 GB
Too big Get file →
Q5_K_M15.5 GB 16.5 GB
Too big Get file →
Q6_K17.9 GB 18.9 GB
Too big Get file →

For 8‑GB VRAM, use IQ1_S or IQ1_M. For 12‑GB VRAM, IQ2_XS or IQ2_S are suitable. For 16‑GB VRAM, Q3_K_S or Q3_K_M provide a good balance of size and performance. Estimates are approximate.

What GPU do you need for ERNIE 4.5 21B A3B PT?

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: the best quant that fits is IQ1_M (5.0 GB file, about 5.9 GB in use).
  • 8 GB VRAM: the best quant that fits is IQ2_S (6.5 GB file, about 7.5 GB in use).
  • 12 GB VRAM: the best quant that fits is Q3_K_M (10.5 GB file, about 11.4 GB in use).
  • 16 GB VRAM: the best quant that fits is Q5_K_S (15.1 GB file, about 16.0 GB in use).
  • 24 GB VRAM: the best quant that fits is Q6_K (17.9 GB file, about 18.9 GB in use).

Which ERNIE 4.5 21B A3B PT quant should you choose?

Q4_K_M is the usual starting point for ERNIE 4.5 21B A3B PT: it balances file size and quality well. Here it is a 13.2 GB file.

If memory is tight, the smallest file is IQ1_S at 4.5 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

Architectureernie4_5-moe
Parameters21.8B (approximate, mixture of experts)
Active parameters3B
Layers28
Trained context131,072 tokens
KV cache (fp16)56 KB per token

Source

Quantized bymradermacher
FamilyErnie
Licenseapache-2.0
Downloads (HF)747
Last updated2025-07-18

Frequently asked questions about ERNIE 4.5 21B A3B PT

How much VRAM does ERNIE 4.5 21B A3B PT need?

About 14.2 GB at Q4_K_M with an 8K context window. That is the 13.2 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 ERNIE 4.5 21B A3B PT on an 8 GB GPU?

Yes, with IQ2_S: the 6.5 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 ERNIE 4.5 21B A3B PT GGUF?

It depends on the quant: IQ1_S is 4.5 GB, IQ1_M is 5.0 GB, IQ2_XXS is 5.8 GB, IQ2_XS is 6.5 GB, IQ2_S is 6.5 GB, IQ2_M is 7.2 GB and 17 more.

What context length does ERNIE 4.5 21B A3B PT support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.5 GB at 8K and 1.9 GB at 32K.

Where can I download ERNIE 4.5 21B A3B PT GGUF?

The files are on Hugging Face in the mradermacher/ERNIE-4.5-21B-A3B-PT-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 ERNIE 4.5 21B A3B PT 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 or IQ1_M quant, which estimate 5.4–5.9 GB memory at 8k context.

Does this model support image inputs?

No, this is a text‑only MoE model with no vision capability.

What license applies to this model?

The model is released under the Apache‑2.0 license.

Other GGUFs of ERNIE-4.5-21B-A3B-PT

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Ernie models

All Ernie →

Other medium models

All medium models →