---

license: apache-2.0

base_model: Qwen/Qwen3.8-27B

library_name: gguf

tags:

• quantization
• trellis-codec
• paw
• gguf

---

PAW-27B

A ~2.17 bpw trellis-coded quantization of Qwen/Qwen3.8-27B that runs on

one 24 GB GPU with a 256k context at 38–78 tok/s.

| | |

|---|---|

| size | 7.814 GB (this file: PAW-27B.gguf) |

| context | 262,144 tokens on a single RTX 3090 |

| speed | 78.4 tok/s short · 47.8 tok/s at 191k code context (MTP drafter) |

| code | 87.2% HumanEval+, 75.9% MBPP+ |

| general | 58.6% MMLU-Pro, 94% GSM8K |

Requirements — read this first

This is not a standard GGUF. It uses the PAW trellis format (paw-dense

arch) and requires the llama-paw

fork. Stock llama.cpp cannot read this model.

How it compares

Against two same-family IQ2XXS GGUFs, measured by us on the same machine with

byte-identical harnesses:

| | size | MMLU-Pro (500) | GSM8K (100) | HumanEval+ | MBPP+ | IFBench-64 strict |

|---|---:|---:|---:|---:|---:|---:|

| PAW-27B | 7.814 GB | 58.6% | 94% | 87.2% | 75.9% | 23.44% |

| Unsloth IQ2XXS | 7.266 GB | 46.2% | 93% | 84.1% | 72.8% | 14.06% |

| AtomicChat AD-IQ2XXS | 8.977 GB | 53.0% | 94% | 73.8% | 71.7% | 20.31% |

PAW-27B leads MMLU-Pro, HumanEval+, MBPP+ and strict IFBench. Unsloth's file is

0.548 GB smaller and gives up 12.4 points of MMLU-Pro. AtomicChat is larger

than PAW-27B and does not recover the coding gap.

Honest summary

These are peer comparisons, not retention numbers. No HumanEval, MBPP or

GSM8K figure exists for Qwen/Qwen3.8-27B at any precision, so there is no

published denominator to claim retention against on those axes, and we are not

going to invent one. What the table shows is that at roughly the same file

size this quantization is ahead of the two obvious alternatives — not how much

of the unquantized parent survives.

The references that do exist for this parent are IFBench (79.5 unquantized,

official Qwen) and LiveCodeBench v6 (90.3). The IFBench figure here is a

64-item defect sentinel, not the 300-item benchmark, so it is directional

only and not comparable to the official 79.5.

Serving

See SERVING.md for the full configuration. Short version:

• requires llama-paw, not stock llama.cpp
• speculative decoding with the parent's own MTP head

(--spec-type draft-mtp) — 22.7 → 78.4 tok/s, lossless

• -ub 2048 is a hard ceiling at -c 262144, not a preference
• KV at q8_0 is free; -fa on is required for it
• for agent/tool use, do not send reasoning back across turns — see the

loop limitation below

Limitations

• Agent clients loop unless reasoning is dropped from history. The model

copies its own prior <think> block verbatim and then repeats the action it

went with. Serve with --no-reasoning-preserve and configure the client

not to return reasoning_content. Details and measurements in

SERVING.md.

• Unbounded thinking returns empty answers. Use

--reasoning-budget 2048, not -1.

• Rare-glyph copy defect. The model cannot emit ➞ (U+279E) and aborts

generation on prompts containing it; in this fork the server returns HTTP

500. Eight HumanEval prompts contain it and it reproduces deterministically

under greedy decoding. This is a model defect, not a harness artifact, and

the affected tasks are scored as failures in the numbers above.

• No parent denominator on HumanEval/MBPP/GSM8K (see above).
• IFBench here is a 64-item sentinel, not the full benchmark.
• Long-context quality is unmeasured; only speed and footprint were measured

at 191k–256k.

Integrity

sha256 f5ec4b409de07f43bbe7bb5d9b3215181c53f5a26876821f732839a2ea3ab8cf  PAW-27B.gguf

License

Apache 2.0, inherited from the base model Qwen/Qwen3.8-27B.