QW

Qwen3.5 REAP 262B A17B abliterated

QwenGGUF by RobinsonLabsUpdated 2026-09-12
Download on Hugging Face
Parameters260B
Context256K
Layers60
Architectureqwen35moe
Smallest quant69.3 GB
Quick answer

Qwen3.5 REAP 262B A17B abliterated needs about 160.1 GB of memory at Q4_K_M (a 158.6 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 10 quants, from 69.3 GB (IQ2_XXS) to 185.7 GB (Q5_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 4.0 GB.

Qwen3.5 REAP 262B A17B abliterated – a 262B‑parameter MoE model with 17B active parameters per token

This model is a community‑built variant of Qwen3.5‑397B‑A17B, pruned with REAP to reduce total parameters to 262B while keeping 17B active per token. It retains vision‑language capabilities and is released under Apache‑2.0. The model is quantized into several GGUF rungs.

Good at

  • Large‑context text generation
  • Vision‑language tasks
  • Low‑refusal content handling

Limits

  • Requires 8‑k context for memory estimates
  • Large VRAM needed for higher‑bit quants
  • No official fine‑tuning support

Best for

  • Developers needing a 262B MoE model with reduced size
  • Researchers testing vision‑language pruning
  • Applications needing controlled refusal behavior

Who it’s for. Engineers and researchers who want a 262B‑parameter Qwen3.5 variant with 17B active parameters per token and who can run it on hardware that supports large VRAM and 8‑k context. The model is suitable for those who need vision‑language support and are comfortable with community‑maintained builds.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_XXS69.3 GB 70.8 GB
Too big Get file →
IQ2_XS77.1 GB 78.6 GB
Too big Get file →
IQ2_M85.7 GB 87.2 GB
Too big Get file →
IQ3_XS107.3 GB 108.8 GB
Too big Get file →
IQ3_M114.6 GB 116.1 GB
Too big Get file →
Q3_K_M125.0 GB 126.5 GB
Too big Get file →
IQ4_XS139.7 GB 141.2 GB
Too big Get file →
Q4_K_S148.7 GB 150.2 GB
Too big Get file →
Q4_K_M158.6 GB 160.1 GB
Too big Get file →
Q5_K_M185.7 GB 187.2 GB
Too big Get file →

For 8‑GB VRAM, use IQ2_XXS (69.26 GB file, ~70.8 GB memory at 8 k context). For 16‑GB VRAM, IQ3_XS (107.28 GB file, ~108.8 GB memory) or Q3_K_M (125 GB file, ~126.5 GB memory). Higher‑bit quants require proportionally more VRAM; estimates are approximate.

What GPU do you need for Qwen3.5 REAP 262B A17B abliterated?

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 Qwen3.5 REAP 262B A17B abliterated quant should you choose?

Q4_K_M is the usual starting point for Qwen3.5 REAP 262B A17B abliterated: it balances file size and quality well. Here it is a 158.6 GB file.

If memory is tight, the smallest file is IQ2_XXS at 69.3 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

Architectureqwen35moe
Parameters260B (approximate, mixture of experts)
Active parameters17B
Layers60
Trained context262,144 tokens
KV cache (fp16)120 KB per token

Source

Quantized byRobinsonLabs
FamilyQwen
Licenseapache-2.0
Downloads (HF)688
Last updated2026-09-12

Frequently asked questions about Qwen3.5 REAP 262B A17B abliterated

How much VRAM does Qwen3.5 REAP 262B A17B abliterated need?

About 160.1 GB at Q4_K_M with an 8K context window. That is the 158.6 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 Qwen3.5 REAP 262B A17B abliterated on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 70.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Qwen3.5 REAP 262B A17B abliterated GGUF?

It depends on the quant: IQ2_XXS is 69.3 GB, IQ2_XS is 77.1 GB, IQ2_M is 85.7 GB, IQ3_XS is 107.3 GB, IQ3_M is 114.6 GB, Q3_K_M is 125.0 GB and 4 more.

What context length does Qwen3.5 REAP 262B A17B abliterated support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.0 GB at 8K and 4.0 GB at 32K.

Where can I download Qwen3.5 REAP 262B A17B abliterated GGUF?

The files are on Hugging Face in the RobinsonLabs/Qwen3.5-REAP-262B-A17B-abliterated-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 Qwen3.5 REAP 262B A17B abliterated 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?

Only the lowest‑bit quant IQ2_XXS can fit, but it still requires about 70 GB of VRAM at 8 k context, so 8 GB is insufficient.

Does this model support vision tasks?

Yes, the vision‑language modality is retained after REAP pruning.

Is the model safe for all audiences?

It is tagged not‑for‑all‑audiences; the hard‑refusal reflex is reduced but child‑safety guardrails remain.

Other GGUFs of Qwen3.5-REAP-262B-A17B

Similar models

Browse all →

More Qwen models

All Qwen →

Other large models

All large models →