GL

Glm 5.3 Flash Reap50 Fp8 Pass2

GLMGGUF by patrickbdevaneyUpdated 2026-10-26
Download on Hugging Face
Parameters165B
Context1024K
Layers46
Architectureglm5-next
Smallest quant72.1 GB
Quick answer

Glm 5.3 Flash Reap50 Fp8 Pass2 needs about 149.2 GB of memory at Q4_K_M (a 99.3 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 5 quants, from 72.1 GB (IQ3_M) to 99.3 GB (Q4_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 197.6 GB.

GLM‑5.3‑Flash REAP‑50 50% expert‑pruned 160B parameter model

GLM‑5.3‑Flash REAP‑50 is a 160 billion‑parameter, multimodal model based on the GLM‑5.3‑Flash architecture. It has been 50 % expert‑pruned and quantised to 4‑bit GGUF weights. The model supports a maximum context length of 1,048,576 tokens and is released under the MIT license.

Good at

  • Multimodal vision and text tasks
  • Large‑context reasoning up to 1M tokens
  • Efficient inference with 4‑bit quantisation

Limits

  • Requires 8 k context for memory estimates
  • No explicit vision support beyond detection
  • Large VRAM needed for higher‑bit quants

Best for

  • Chatbots needing long‑context understanding
  • Vision‑enabled content generation
  • Research prototyping with multimodal models

Who it’s for. Developers and researchers who need a multimodal LLM with long‑context capability and who can manage the memory requirements of 4‑bit GGUF files. The model is suitable for local deployment using llama.cpp or other supported frameworks.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ3_M72.1 GB 122.0 GB
Too big Get file →
Q3_K_M78.8 GB 128.7 GB
Too big Get file →
IQ4_XS88.0 GB 137.9 GB
Too big Get file →
Q4_K_S93.5 GB 143.4 GB
Too big Get file →
Q4_K_M99.3 GB 149.2 GB
Too big Get file →

For 8 GB VRAM use IQ3_M (72.13 GB file, ~122 GB VRAM at 8 k context). For 12 GB VRAM use Q3_K_M (78.78 GB file, ~129 GB VRAM). For 16 GB VRAM use Q4_K_M (99.33 GB file, ~149 GB VRAM). Estimates are approximate.

What GPU do you need for Glm 5.3 Flash Reap50 Fp8 Pass2?

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: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: no quant fits at an 8K context.

Which Glm 5.3 Flash Reap50 Fp8 Pass2 quant should you choose?

Q4_K_M is the usual starting point for Glm 5.3 Flash Reap50 Fp8 Pass2: it balances file size and quality well. Here it is a 99.3 GB file.

If memory is tight, the smallest file is IQ3_M at 72.1 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureglm5-next
Parameters165B (approximate, mixture of experts)
Layers46
Trained context1,048,576 tokens
KV cache (fp16)5888 KB per token

Source

Quantized bypatrickbdevaney
FamilyGLM
Licensemit
Downloads (HF)33,475
Last updated2026-10-26

Frequently asked questions about Glm 5.3 Flash Reap50 Fp8 Pass2

How much VRAM does Glm 5.3 Flash Reap50 Fp8 Pass2 need?

About 149.2 GB at Q4_K_M with an 8K context window. That is the 99.3 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 Glm 5.3 Flash Reap50 Fp8 Pass2 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 122.0 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Glm 5.3 Flash Reap50 Fp8 Pass2 GGUF?

It depends on the quant: IQ3_M is 72.1 GB, Q3_K_M is 78.8 GB, IQ4_XS is 88.0 GB, Q4_K_S is 93.5 GB, Q4_K_M is 99.3 GB.

What context length does Glm 5.3 Flash Reap50 Fp8 Pass2 support?

The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 49.4 GB at 8K and 197.6 GB at 32K.

Where can I download Glm 5.3 Flash Reap50 Fp8 Pass2 GGUF?

The files are on Hugging Face in the patrickbdevaney/GLM-5.3-Flash-REAP50-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 Glm 5.3 Flash Reap50 Fp8 Pass2 use?

Hugging Face lists the license as mit. 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 IQ3_M quantisation; it requires about 122 GB VRAM at 8 k context, so 8 GB is insufficient.

Does this model support vision?

Yes, the model has detected vision abilities, but no explicit vision tokenisation details are provided.

What license does it use?

The model is released under the MIT license.

Other GGUFs of GLM-5.3-Flash

Similar models

Browse all →

More from patrickbdevaney

All by patrickbdevaney →

More GLM models

All GLM →

Other large models

All large models →