0xSero/GLM-5.1-444B-A14B-REAP-GGUF

BROKEN - Do Not Use

This model has repetition degeneration. Use the 25% pruned version instead.

Use this instead: 0xSero/GLM-5.1-555B-A14B-REAP-GGUF

---

What is wrong with this model?

This is a Q4_K_M GGUF of the 40% expert-pruned GLM-5.1 (154/256 experts retained). It suffers from repetition degeneration - the model enters infinite loops when generating code, structured output, or any long-form content requiring syntactic templates.

Measured degeneration rates:

• 29% overall (13/45 probes degenerate in fuzz testing)
• 40% of code generation tasks loop (red-black trees, chess engines, regex, B-trees)
• 75% of structured output tasks loop (comparison tables, API specs, enum lists)
• 18% of Terminal-Bench probes loop (9/50)
• 30% of SWE-bench Pro probes loop (12/40)

Root cause:

Removing 40% of experts (102 per layer) exceeds the model's tolerance for expert pruning. The remaining 154 experts cannot cover the full routing distribution needed for coherent long-form generation. The degeneration compounds over sequence length - short outputs (<512 tokens) work fine, but anything over ~600-1000 words risks entering a repetition loop.

The fix:

The 25% pruned variant (192/256 experts, 555B) completely eliminates repetition loops while maintaining competitive quality:

• 0/220 benchmark probes had repetition loops
• Terminal-Bench: 88% proxy pass rate
• SWE-Pro: 66% proxy pass rate

Use 0xSero/GLM-5.1-555B-A14B-REAP-GGUF instead.

Sponsors

Thank you for the kind sponsors, wouldn't be possible without them:

• Nvidia
• TNG Technology
• Lambda
• Prime Intellect
• HotAisle

GLM-5.1 REAP Family — Hardware Compatibility

All variants in this family are REAP-pruned (2510.13999) descendants of zai-org/GLM-5.1 (original: 744B params, 256 experts/MoE layer, 40B activated/token). Pick a variant based on your GPU architecture and available VRAM.

Quick picker

You have	Use
8× H100/H200 80GB (Hopper, sm_90)	GLM-5.1-555B-A14B-REAP-GPTQ-W4A16 or GLM-5.1-555B-A14B-REAP-NVFP4 (NVFP4 on Hopper via modelopt_fp4 + triton path)
4× RTX PRO 6000 Blackwell Workstation 96GB (sm_120)	GLM-5.1-478B-A42B-REAP-NVFP4 (further-pruned 160-expert, 200k ctx) — this is the Blackwell Workstation reference config
4× B200 180GB (sm_100)	GLM-5.1-478B-A42B-REAP-NVFP4 or GLM-5.1-555B-A14B-REAP-NVFP4
8× B200 / Blackwell datacenter	GLM-5.1-555B-A14B-REAP-NVFP4 (192-expert, upstream's reference config with flashinfer + b12x backends)
8× A100 80GB (Ampere, sm_80)	GLM-5.1-444B-A14B-REAP (BF16) or -GPTQ-W4A16
CPU / Apple Silicon / consumer GPU with llama.cpp	GLM-5.1-555B-A14B-REAP-GGUF or GLM-5.1-444B-A14B-REAP-GGUF

Full family

Variant	Format	Size	Experts/layer	Activated/token	Min VRAM (TP)	Inference engine	Best on
GLM-5.1-555B-A14B-REAP	BF16	~1125 GB	192	~14B	8× 141 GB (H200)	sglang / vllm	Hopper
GLM-5.1-444B-A14B-REAP	BF16	~910 GB	154	~14B	8× 114 GB	sglang / vllm	Ampere / Hopper
GLM-5.1-555B-A14B-REAP-NVFP4	NVFP4 (4-bit)	~320 GB	192	~14B	4× 80 GB (B200), 8× 48 GB	sglang --quantization modelopt_fp4	Blackwell (native); Hopper (triton path)
GLM-5.1-478B-A42B-REAP-NVFP4	NVFP4 (4-bit)	~285 GB	160	~42B	4× 80 GB Blackwell	sglang --quantization modelopt_fp4	4× RTX PRO 6000 Blackwell @ 200k ctx
GLM-5.1-555B-A14B-REAP-GPTQ-W4A16	GPTQ W4A16	~297 GB	192	~14B	4× 80 GB	vllm / sglang --quantization gptq_marlin	Hopper (best), works on Ampere
GLM-5.1-555B-A14B-REAP-GGUF	GGUF (Q2–Q8)	~348 GB	192	~14B	Varies by quant	llama.cpp	CPU / Apple / consumer CUDA
GLM-5.1-444B-A14B-REAP-GGUF	GGUF (Q2–Q8)	~283 GB	154	~14B	Varies by quant	llama.cpp	CPU / Apple / consumer CUDA

Notes

• NVFP4 on Hopper (H100/H200): supported from sglang 25.10 / 0.5.10+ (NVIDIA SGLang release notes); native Blackwell tensor-core FP4 still gives better throughput.
• NVFP4 on B200 / Blackwell datacenter (sm_100): use flashinfer attention + b12x or flashinfer MoE backends — this is the recipe in the original 555B-A14B-REAP-NVFP4 card.
• NVFP4 on Blackwell Workstation (sm_120): use --attention-backend triton (not flashinfer — PCIe P2P atomics unavailable on the consumer board), --moe-runner-backend cutlass, --fp4-gemm-backend flashinfer_cudnn. See the GLM-5.1-478B-A42B-REAP-NVFP4 card for the full 200k-ctx replication guide.
• GPTQ-W4A16 vs NVFP4: same bit depth, different hardware path. NVFP4 has native Blackwell support and per-16 fp8 scales; GPTQ is group-quantized int4 with broader engine support.
• REAP expert count variants (555B/444B): different expert-retention ratios from the same base; 555B keeps more experts (higher quality ceiling), 444B trades quality for 20% less VRAM.
• Why NVFP4-478B-A42B-REAP is different: it's double-pruned (256 → 192 → 160 experts), optimized for a specific Blackwell Workstation 4×96GB target at 200k context. The A42B suffix reflects measured activated params/token on the 160-expert MoE, not the REAP branding convention of the sibling variants.

Pointer to active inference recipe

See GLM-5.1-478B-A42B-REAP-NVFP4 README for the full Blackwell Workstation replication guide (exact software pins, NSA patch, launch flags, measured 200k-ctx perf, sampling recommendations). Most of the sglang flags carry over to other NVFP4 variants on other hardware.

Citation

@misc{lasby2025reap,
  title={REAP the Experts: Why Pruning Prevails for One-Shot MoE compression},
  author={Mike Lasby and Ivan Lazarevich and Nish Sinnadurai and Sean Lie and Yani Ioannou and Vithursan Thangarasa},
  year={2025},
  eprint={2510.13999},
  archivePrefix={arXiv},
}