Recipe
qwen3-8-27b-uncensored-int4-autoround-gptq-int4-intel-arc-pro-b70-32gb-vllm-tp1
qwen3-8-27b-uncensored-int4-autoround-gptq-int4-intel-arc-pro-b70-32gb-vllm-tp1Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.
Record
- Status
- candidate
- Source
- localmaxxing
- Engine
- vllm
- Engine version
- 0.27.2rc1.dev78+g8e6d8e4f6 XPU (custom-patched: host-staged TP2 collectives + INT8 W8A8 lm_head + MTP wiring)
- Accelerators
- 1
- Tensor parallel
- 1
- Context tokens
- 2,048
- Max concurrency
- 1
- chat
- unknown
- reasoning
- unknown
- tools
- unknown
- vision
- unknown
Hugging Face model card
Identity
https://huggingface.co/johannrplaster/Qwen3.8-27B-Uncensored-int4-AutoRound- Repository
- johannrplaster/Qwen3.8-27B-Uncensored-int4-AutoRound
- Status
- known
- Link type
- Exact Hub repository
Public Hugging Face repository confirmed by the Hub API.
Candidate: useful compatibility or speed evidence. The registry does not offer Run until promotion requirements are met.
Observed configuration
vllm
Evidence only · candidate · reference
Candidate evidence — not a Run contract
Observed source tokens
Mechanical split of the source command. Unverified against the engine CLI. The registry does not offer Run for this recipe.
vllmservejohannrplaster/Qwen3.8-27B-Uncensored-int4-AutoRound--host0.0.0.0--port19622--served-model-nameQwen3.8-27B-UNC-G128-AR--tensor-parallel-size1--dtypefloat16--max-model-len8192--max-num-seqs1--max-num-batched-tokens8192--gpu-memory-utilization0.88--quantizationgptq--speculative-config{"method":"qwen3_5_mtp","num_speculative_tokens":3}
| Flag | Value |
|---|---|
--host | 0.0.0.0 |
--port | 19622 |
--served-model-name | Qwen3.8-27B-UNC-G128-AR |
--tensor-parallel-size | 1 |
--dtype | float16 |
--max-model-len | 8192 |
--max-num-seqs | 1 |
--max-num-batched-tokens | 8192 |
--gpu-memory-utilization | 0.88 |
--quantization | gptq |
--speculative-config | {"method":"qwen3_5_mtp","num_speculative_tokens":3} |
Environment
| Variable | Value |
|---|---|
CCL_TOPO_P2P_ACCESS | 0 |
ONEAPI_DEVICE_SELECTOR | level_zero:0 |
VLLM_XPU_HOST_STAGED_COLLECTIVES | 1 |
VLLM_XPU_LM_HEAD_INT8 | 1 |
ZE_AFFINITY_MASK | 0 |
Measured speed
| Concurrency | Context | Prefill | Decode | TTFT ms | Status | Sweep |
|---|---|---|---|---|---|---|
| — | 2,048 | 1,782.1 | 97.8 | 356.3 | observed | qwen3-8-27b-uncensored-int4-autoround-gptq-int4-intel-arc-pro-b70-32gb-vllm-tp1-sweep |
Remaining fields
Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.
- hardware count
- 1
- hardware id
- intel-arc-pro-b70-32gb
- id
- qwen3-8-27b-uncensored-int4-autoround-gptq-int4-intel-arc-pro-b70-32gb-vllm-tp1
- model instance id
- johannrplaster-qwen3-8-27b-uncensored-int4-autoround--gptq-int4
- recipe source
- localmaxxing
- schema version
- local-ai-registry/v1
- speed sweep ids
- qwen3-8-27b-uncensored-int4-autoround-gptq-int4-intel-arc-pro-b70-32gb-vllm-tp1-sweep
- status
- candidate
capabilities
engine
- name
- vllm
- version
- 0.27.2rc1.dev78+g8e6d8e4f6 XPU (custom-patched: host-staged TP2 collectives + INT8 W8A8 lm_head + MTP wiring)
serving
- max concurrency
- 1
- max context tokens
- 2,048
- tensor parallel
- 1
Provenance & metadata (3)
facts
- capabilities.chat · provenance · captured at
- 2026-08-30T09:10:02Z
capabilities.chat · reason capability-not-verified
capabilities.chat · state unknown
- capabilities.reasoning · provenance · captured at
- 2026-08-30T09:10:02Z
capabilities.reasoning · reason capability-not-verified
capabilities.reasoning · state unknown
- capabilities.tools · provenance · captured at
- 2026-08-30T09:10:02Z
capabilities.tools · reason capability-not-verified
capabilities.tools · state unknown
- capabilities.vision · provenance · captured at
- 2026-08-30T09:10:02Z
capabilities.vision · reason capability-not-verified
capabilities.vision · state unknown
- engine.graph mode · provenance · captured at
- 2026-08-30T09:10:02Z
engine.graph mode · reason runtime-detail-not-published
engine.graph mode · state unknown
- serving.kv cache tokens · provenance · captured at
- 2026-08-30T09:10:02Z
serving.kv cache tokens · reason kv-cache-capacity-not-published
serving.kv cache tokens · state unknown
- serving.max concurrency · provenance · captured at
- 2026-08-31T23:03:15Z
serving.max concurrency · reason server-capacity-derived-from-source-evidence
serving.max concurrency · state known
metadata
- localmaxxing · backend
- xpu
- localmaxxing · batch size
- 1
- localmaxxing · hardware label
- Intel Arc Pro B70
- localmaxxing · notes
- johannrplaster/Qwen3.8-27B-Uncensored-int4-AutoRound GPTQ-Int4 served via vLLM Marlin after auto-round->gptq config flip; GDN linear-attention in_proj_a/b kept FP16 (quality-preserving); load-time INT8 W8A8 lm_head (per-channel RTN). 1x Intel Arc Pro B70, MTP3 speculative decode, FP16 KV, greedy. Strict cold suite 66.8 tok/s, quality gates pass.
- localmaxxing · observed command
- VLLM_XPU_LM_HEAD_INT8=1 VLLM_XPU_HOST_STAGED_COLLECTIVES=1 CCL_TOPO_P2P_ACCESS=0 ZE_AFFINITY_MASK=0 ONEAPI_DEVICE_SELECTOR=level_zero:0 vllm serve johannrplaster/Qwen3.8-27B-Uncensored-int4-AutoRound --host 0.0.0.0 --port 19622 --served-model-name Qwen3.8-27B-UNC-G128-AR --tensor-parallel-size 1 --dtype float16 --max-model-len 8192 --max-num-seqs 1 --max-num-batched-tokens 8192 --gpu-memory-utilization 0.88 --quantization gptq --speculative-config '{"method":"qwen3_5_mtp","num_speculative_tokens":3}'
- localmaxxing · run id
- cmt0vrxjk0fvpms01n4i9cmns
- localmaxxing · tokenized · arguments
- vllm, serve, johannrplaster/Qwen3.8-27B-Uncensored-int4-AutoRound, --host, 0.0.0.0, --port, 19622, --served-model-name, Qwen3.8-27B-UNC-G128-AR, --tensor-parallel-size, 1, --dtype, float16, --max-model-len, 8192, --max-num-seqs, 1, --max-num-batched-tokens, 8192, --gpu-memory-utilization, 0.88, --quantization, gptq, --speculative-config, {"method":"qwen3_5_mtp","num_speculative_tokens":3}
- localmaxxing · tokenized · environment · CCL TOPO P2P ACCESS
- 0
- localmaxxing · tokenized · environment · ONEAPI DEVICE SELECTOR
- level_zero:0
- localmaxxing · tokenized · environment · VLLM XPU HOST STAGED COLLECTIVES
- 1
- localmaxxing · tokenized · environment · VLLM XPU LM HEAD INT8
- 1
- localmaxxing · tokenized · environment · ZE AFFINITY MASK
- 0
- localmaxxing · tokenized · fidelity
- faithful
provenance
- captured at
- 2026-08-30T09:10:02Z
sources
| captured at | kind | url |
|---|---|---|
| 2026-08-30T09:10:02Z | normalized-recipe | www.localmaxxing.com/en/runs/cmt0vrxjk0fvpms01n4i9cmns ↗ |