Recipe

qwen3-6-27b-nvfp4-rtx-pro-6000-blackwell-96gb-vllm-tp1

qwen3-6-27b-nvfp4-rtx-pro-6000-blackwell-96gb-vllm-tp1

Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.

Record

Status
candidate
Source
localmaxxing
Engine
vllm
Engine version
0.26.0
Accelerators
1
Tensor parallel
1
Context tokens
2,048
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/unsloth/Qwen3.6-27B-NVFP4
Repository
unsloth/Qwen3.6-27B-NVFP4
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

Endpoint
http://127.0.0.1:8000
Served model
qwen3.6-27b
Source
https://www.localmaxxing.com/en/runs/cmrhgyn5d00oumj017fz6ia7p

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
12,04814052.4observedqwen3-6-27b-nvfp4-rtx-pro-6000-blackwell-96gb-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
rtx-pro-6000-blackwell-96gb
id
qwen3-6-27b-nvfp4-rtx-pro-6000-blackwell-96gb-vllm-tp1
model instance id
unsloth-qwen3-6-27b-nvfp4--nvfp4
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
qwen3-6-27b-nvfp4-rtx-pro-6000-blackwell-96gb-vllm-tp1-sweep
status
candidate

capabilities

draft launch

accelerator backend
nvidia
arguments
serve, --model, unsloth/Qwen3.6-27B-NVFP4, --tensor-parallel-size, 1, --host, 0.0.0.0, --port, 8000, --max-model-len, 2048
container port
8,000
entrypoint
vllm
host port
8,000
image
vllm/vllm-openai@sha256:0a51ea5b4ae2dc5d81890e5173f54203d2a3ae0cfffe51b8fd2afd4391bfd967
ipc
host
kind
docker
shm size
16g
synthesized · generated at
2026-08-31T22:12:17Z
synthesized · image provenance
deepseek-fp8-rtx-pro-6000-blackwell-96gb-vllm-tp1
synthesized · template
vllm-openai-v1

mounts

read onlytarget
No/root/.cache/huggingface

engine

name
vllm
version
0.26.0

serving

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
draft launch.entrypoint · provenance · captured at
2026-09-01T01:41:26Z
draft launch.entrypoint · reason linux-amd64-container-config-entrypoint
draft launch.entrypoint · state known
engine.graph mode · provenance · captured at
2026-08-30T09:10:02Z
engine.graph mode · reason runtime-detail-not-published
engine.graph mode · state unknown
engine.version · provenance · captured at
2026-08-31T23:38:26Z
engine.version · reason copied-from-synthesized-image-provenance
engine.version · state known
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-not-evidenced
serving.max concurrency · state unknown

metadata

localmaxxing · backend
cuda
localmaxxing · batch size
1
localmaxxing · hardware label
RTX PRO 6000 Blackwell
localmaxxing · notes
Production co-resident config (ComfyUI shares the GPU) - not a tuned bench rig. vllm serve unsloth/Qwen3.6-27B-NVFP4 \ --served-model-name qwen3.6-27b --trust-remote-code \ --gpu-memory-utilization 0.40 --max-num-seqs 64 --max-model-len 262144 \ --kv-cache-dtype fp8_e4m3 --reasoning-parser qwen3 \ --tool-call-parser qwen3_xml --enable-auto-tool-choice \ --speculative-config {"method":"qwen3_5_mtp","num_speculative_tokens":5} This checkpoint carries BOTH the MTP draft head and the vision tower, so one engine gets spec-decode speed AND image support (verified on vLLM 0.24.0). num_speculative_tokens sweep, single-stream, same rig: 3 -> 137.1 tok/s | 5 -> 140.1 tok/s | 8 -> 130.2 tok/s mean accepted per draft: 2.46 at depth 5 vs 2.82 at depth 8 (35% rate) - knee is at 5. TTFT measured client-side as time to first streamed delta (the qwen3 reasoning parser streams into the reasoning field, which hides first-token timing from generic harnesses). Power: nvidia-smi 1Hz peak during decode. Driver 595-server-open, CUDA 13.1.
localmaxxing · observed command
# Remote endpoint: http://127.0.0.1:8000 servedModel: qwen3.6-27b
localmaxxing · run id
cmrhgyn5d00oumj017fz6ia7p
localmaxxing · tokenized · endpoint
127.0.0.1
localmaxxing · tokenized · fidelity
faithful
localmaxxing · tokenized · served model name
qwen3.6-27b

provenance

captured at
2026-08-30T09:10:02Z

sources

captured atkindurl
2026-08-30T09:10:02Znormalized-recipewww.localmaxxing.com/en/runs/cmrhgyn5d00oumj017fz6ia7p