GL

cerebras/GLM-4.7-Flash-REAP-23B-A3B

GLMGGUF by mradermacherUpdated 2026-10-30
Download on Hugging Face
Parameters23B
Context198K
Layers47
Architecturedeepseek2
Smallest quant8.5 GB
Quick answer

cerebras/GLM-4.7-Flash-REAP-23B-A3B needs about 15.3 GB of memory at Q4_K_M (a 13.9 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 11 quants, from 8.5 GB (Q2_K) to 24.5 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 3.4 GB.

GLM‑4.7‑Flash‑REAP‑23B‑A3B – a 23B sparse Mixture‑of‑Experts model

GLM‑4.7‑Flash‑REAP‑23B‑A3B is a 23B parameter model compressed from 30B using REAP pruning. It retains near‑lossless performance on code generation, agentic workflows, and function calling. The model is a sparse MoE with 48 experts, 4 activated per token.

Good at

  • Reasoning
  • Tools usage
  • Code generation

Limits

  • Requires 8k context for memory estimates
  • No vision support
  • MIT license only

Best for

  • Local deployment on limited GPU
  • Academic research on sparse MoE
  • Function‑calling agents

Who it’s for. Researchers and developers needing a memory‑efficient GLM variant for local inference. The model runs on standard GPU setups without custom patches.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K8.5 GB 9.9 GB
Too big Get file →
Q3_K_S10.0 GB 11.4 GB
Too big Get file →
Q3_K_M11.1 GB 12.4 GB
Too big Get file →
Q3_K_L12.0 GB 13.3 GB
Too big Get file →
IQ4_XS12.4 GB 13.8 GB
Too big Get file →
Q4_K_S13.1 GB 14.5 GB
Too big Get file →
Q4_K_M13.9 GB 15.3 GB
Too big Get file →
Q5_K_S15.9 GB 17.2 GB
Too big Get file →
Q5_K_M16.3 GB 17.7 GB
Too big Get file →
Q6_K18.9 GB 20.2 GB
Too big Get file →
Q8_024.5 GB 25.8 GB
Too big Get file →

For 8 GB VRAM use Q2_K (8.52 GB file, ~9.9 GB memory). For 12 GB VRAM use Q3_K_M (11.08 GB file, ~12.4 GB memory). For 16 GB VRAM use Q4_K_M (13.93 GB file, ~15.3 GB memory).

What GPU do you need for cerebras/GLM-4.7-Flash-REAP-23B-A3B?

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 Q3_K_S (10.0 GB file, about 11.4 GB in use).
  • 16 GB VRAM: the best quant that fits is Q4_K_M (13.9 GB file, about 15.3 GB in use).
  • 24 GB VRAM: the best quant that fits is Q6_K (18.9 GB file, about 20.2 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (24.5 GB file, about 25.8 GB in use).

Which cerebras/GLM-4.7-Flash-REAP-23B-A3B quant should you choose?

Q4_K_M is the usual starting point for cerebras/GLM-4.7-Flash-REAP-23B-A3B: it balances file size and quality well. Here it is a 13.9 GB file.

The Q8_0 file is 24.5 GB, about 1.8 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 Q2_K at 8.5 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

Architecturedeepseek2
Parameters23B (approximate, mixture of experts)
Active parameters3B
Layers47
Trained context202,752 tokens
KV cache (fp16)100 KB per token

Source

Quantized bymradermacher
FamilyGLM
Licensemit
Downloads (HF)940
Last updated2026-10-30

Frequently asked questions about cerebras/GLM-4.7-Flash-REAP-23B-A3B

How much VRAM does cerebras/GLM-4.7-Flash-REAP-23B-A3B need?

About 15.3 GB at Q4_K_M with an 8K context window. That is the 13.9 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 cerebras/GLM-4.7-Flash-REAP-23B-A3B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 9.9 GB, so you would need a larger GPU or CPU offloading.

What is the file size of cerebras/GLM-4.7-Flash-REAP-23B-A3B GGUF?

It depends on the quant: Q2_K is 8.5 GB, Q3_K_S is 10.0 GB, Q3_K_M is 11.1 GB, Q3_K_L is 12.0 GB, IQ4_XS is 12.4 GB, Q4_K_S is 13.1 GB and 5 more.

What context length does cerebras/GLM-4.7-Flash-REAP-23B-A3B support?

The GGUF header lists a trained context length of 202,752 tokens (198K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.4 GB at 32K.

Where can I download cerebras/GLM-4.7-Flash-REAP-23B-A3B GGUF?

The files are on Hugging Face in the mradermacher/GLM-4.7-Flash-REAP-23B-A3B-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 cerebras/GLM-4.7-Flash-REAP-23B-A3B 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?

Yes, the Q2_K quant uses 8.52 GB file size and estimates 9.9 GB memory at 8k context, which fits an 8 GB VRAM setup with some overhead.

Does this model support vision tasks?

No, the model has no vision capabilities.

What license applies to this model?

The model is released under the MIT license.

Other GGUFs of GLM-4.7-Flash-REAP-23B-A3B

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More GLM models

All GLM →

Other medium models

All medium models →