Recipe
qwen3-6-35b-a3b-nvfp4-dgx-spark-gb10-128gb-vllm-tp1
qwen3-6-35b-a3b-nvfp4-dgx-spark-gb10-128gb-vllm-tp1Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.
Record
- Status
- candidate
- Source
- localmaxxing
- Engine
- vllm
- Engine version
- 0.23.1rc1.dev272+gb7abab517.d20260717
- Accelerators
- 1
- Tensor parallel
- 1
- Context tokens
- 262,144
- Max concurrency
- 4
- chat
- unknown
- reasoning
- unknown
- tools
- unknown
- vision
- unknown
Hugging Face model card
Identity
https://huggingface.co/nvidia/Qwen3.6-35B-A3B-NVFP4- Repository
- nvidia/Qwen3.6-35B-A3B-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
Observed source tokens
Mechanical split of the source command. Unverified against the engine CLI. The registry does not offer Run for this recipe.
vllmservenvidia/Qwen3.6-35B-A3B-NVFP4--served-model-nameqwen36-35b-a3b-nvfp4--tensor-parallel-size1--trust-remote-code--quantizationmodelopt--kv-cache-dtypefp8--attention-backendflashinfer--moe-backendmarlin--gpu-memory-utilization0.4--max-model-len262144--max-num-seqs4--max-num-batched-tokens8192--enable-chunked-prefill--async-scheduling--enable-prefix-caching--speculative-config{"method":"mtp","num_speculative_tokens":3,"moe_backend":"triton"}--load-formatfastsafetensors--reasoning-parserqwen3--tool-call-parserqwen3_xml--enable-auto-tool-choice
| Flag | Value |
|---|---|
--served-model-name | qwen36-35b-a3b-nvfp4 |
--tensor-parallel-size | 1 |
--quantization | modelopt |
--kv-cache-dtype | fp8 |
--attention-backend | flashinfer |
--moe-backend | marlin |
--gpu-memory-utilization | 0.4 |
--max-model-len | 262144 |
--max-num-seqs | 4 |
--max-num-batched-tokens | 8192 |
--speculative-config | {"method":"mtp","num_speculative_tokens":3,"moe_backend":"triton"} |
--load-format | fastsafetensors |
--reasoning-parser | qwen3 |
--tool-call-parser | qwen3_xml |
Measured speed
| Concurrency | Context | Prefill | Decode | TTFT ms | Status | Sweep |
|---|---|---|---|---|---|---|
| 4 | 262,144 | — | 300.6 | 341 | observed | qwen3-6-35b-a3b-nvfp4-dgx-spark-gb10-128gb-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
- dgx-spark-gb10-128gb
- id
- qwen3-6-35b-a3b-nvfp4-dgx-spark-gb10-128gb-vllm-tp1
- model instance id
- nvidia-qwen3-6-35b-a3b-nvfp4--nvfp4
- recipe source
- localmaxxing
- schema version
- local-ai-registry/v1
- speed sweep ids
- qwen3-6-35b-a3b-nvfp4-dgx-spark-gb10-128gb-vllm-tp1-sweep
- status
- candidate
capabilities
draft launch
- accelerator backend
- nvidia
- arguments
- --model, nvidia/Qwen3.6-35B-A3B-NVFP4, --tensor-parallel-size, 1, --host, 0.0.0.0, --port, 8000, --max-model-len, 262144
- container port
- 8,000
- 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 only | target |
|---|---|
| No | /root/.cache/huggingface |
engine
- name
- vllm
- version
- 0.23.1rc1.dev272+gb7abab517.d20260717
serving
- max concurrency
- 4
- max context tokens
- 262,144
- tensor parallel
- 1
Provenance & metadata (3)
facts
- capabilities.chat · provenance · captured at
- 2026-08-30T09:26:09Z
capabilities.chat · reason capability-not-verified
capabilities.chat · state unknown
- capabilities.reasoning · provenance · captured at
- 2026-08-30T09:26:09Z
capabilities.reasoning · reason capability-not-verified
capabilities.reasoning · state unknown
- capabilities.tools · provenance · captured at
- 2026-08-30T09:26:09Z
capabilities.tools · reason capability-not-verified
capabilities.tools · state unknown
- capabilities.vision · provenance · captured at
- 2026-08-30T09:26:09Z
capabilities.vision · reason capability-not-verified
capabilities.vision · state unknown
- engine.graph mode · provenance · captured at
- 2026-08-30T09:26:09Z
engine.graph mode · reason runtime-detail-not-published
engine.graph mode · state unknown
- serving.kv cache tokens · provenance · captured at
- 2026-08-30T09:26:09Z
serving.kv cache tokens · reason kv-cache-capacity-not-published
serving.kv cache tokens · state unknown
metadata
- localmaxxing · backend
- cuda
- localmaxxing · hardware label
- GB10 Grace Blackwell
- localmaxxing · notes
- Four-concurrent-request text throughput benchmark: 2 warmup waves followed by 5 measured streaming greedy waves; aggregate median reported (291.40-324.31 output tok/s). Each request used 361 prompt and 512 output tokens; batchSize is concurrent requests, while promptTokens/outputTokens are per request. This is 2.42x the same profile's 124.07 tok/s single-stream result. NVIDIA ModelOpt NVFP4 with mixed higher-precision tensors, FP8 KV, Marlin MoE backend, FlashInfer attention, and MTP-3. Prefix caching was enabled, so repeated-prompt TTFT is warm-cache. Metrics use server usage counters and client-observed token timing.
- localmaxxing · observed command
- vllm serve nvidia/Qwen3.6-35B-A3B-NVFP4 --served-model-name qwen36-35b-a3b-nvfp4 --tensor-parallel-size 1 --trust-remote-code --quantization modelopt --kv-cache-dtype fp8 --attention-backend flashinfer --moe-backend marlin --gpu-memory-utilization 0.4 --max-model-len 262144 --max-num-seqs 4 --max-num-batched-tokens 8192 --enable-chunked-prefill --async-scheduling --enable-prefix-caching --speculative-config '{"method":"mtp","num_speculative_tokens":3,"moe_backend":"triton"}' --load-format fastsafetensors --reasoning-parser qwen3 --tool-call-parser qwen3_xml --enable-auto-tool-choice
- localmaxxing · run id
- cms8d1vz600fipf01b9wjtbmm
- localmaxxing · tokenized · arguments
- vllm, serve, nvidia/Qwen3.6-35B-A3B-NVFP4, --served-model-name, qwen36-35b-a3b-nvfp4, --tensor-parallel-size, 1, --trust-remote-code, --quantization, modelopt, --kv-cache-dtype, fp8, --attention-backend, flashinfer, --moe-backend, marlin, --gpu-memory-utilization, 0.4, --max-model-len, 262144, --max-num-seqs, 4, --max-num-batched-tokens, 8192, --enable-chunked-prefill, --async-scheduling, --enable-prefix-caching, --speculative-config, {"method":"mtp","num_speculative_tokens":3,"moe_backend":"triton"}, --load-format, fastsafetensors, --reasoning-parser, qwen3, --tool-call-parser, qwen3_xml, --enable-auto-tool-choice
- localmaxxing · tokenized · fidelity
- faithful
provenance
- captured at
- 2026-08-30T09:26:09Z
sources
| captured at | kind | url |
|---|---|---|
| 2026-08-30T09:26:09Z | normalized-recipe | www.localmaxxing.com/en/runs/cms8d1vz600fipf01b9wjtbmm ↗ |