Recipe
qwen3-6-35b-a3b-nvfp4-rtx-5090-32gb-vllm-tp1-mj82rijd
qwen3-6-35b-a3b-nvfp4-rtx-5090-32gb-vllm-tp1-mj82rijdObserved LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.
Record
- Status
- candidate
- Source
- localmaxxing
- Engine
- vllm
- Engine version
- 0.20.2rc1.dev189+gfb1ac806c
- Accelerators
- 1
- Tensor parallel
- 1
- Context tokens
- 262,144
- Max concurrency
- 8
- chat
- unknown
- reasoning
- unknown
- tools
- unknown
- vision
- unknown
Hugging Face model card
Identity
https://huggingface.co/Infatoshi/Qwen3.6-35B-A3B-NVFP4-FP8- Repository
- Infatoshi/Qwen3.6-35B-A3B-NVFP4-FP8
- 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.
vllmserve/models/Qwen3.6-35B-A3B-NVFP4-FP8--served-model-nameQwen3.6-35B-A3B--reasoning-parserqwen3--enable-auto-tool-choice--tool-call-parserqwen3_coder--max-model-lenauto--limit-mm-per-prompt.video0--max-cudagraph-capture-size32--max-num-seqs8--max-num-batched-tokens4096--gpu-memory-utilization0.940--trust-remote-code--quantizationcompressed-tensors--safetensors-load-strategyprefetch--enable-chunked-prefill--enable-prefix-caching--scheduling-policypriority--language-model-only--attention-backendflash_attn--speculative-config{method:qwen3_next_mtp,num_speculative_tokens:4}
| Flag | Value |
|---|---|
--served-model-name | Qwen3.6-35B-A3B |
--reasoning-parser | qwen3 |
--tool-call-parser | qwen3_coder |
--max-model-len | auto |
--limit-mm-per-prompt.video | 0 |
--max-cudagraph-capture-size | 32 |
--max-num-seqs | 8 |
--max-num-batched-tokens | 4096 |
--gpu-memory-utilization | 0.940 |
--quantization | compressed-tensors |
--safetensors-load-strategy | prefetch |
--scheduling-policy | priority |
--attention-backend | flash_attn |
--speculative-config | {method:qwen3_next_mtp,num_speculative_tokens:4} |
Measured speed
| Concurrency | Context | Prefill | Decode | TTFT ms | Status | Sweep |
|---|---|---|---|---|---|---|
| — | 262,144 | 134.1 | 203.6 | 358 | observed | qwen3-6-35b-a3b-nvfp4-rtx-5090-32gb-vllm-tp1-mj82rijd-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-5090-32gb
- id
- qwen3-6-35b-a3b-nvfp4-rtx-5090-32gb-vllm-tp1-mj82rijd
- model instance id
- infatoshi-qwen3-6-35b-a3b-nvfp4-fp8--nvfp4
- recipe source
- localmaxxing
- schema version
- local-ai-registry/v1
- speed sweep ids
- qwen3-6-35b-a3b-nvfp4-rtx-5090-32gb-vllm-tp1-mj82rijd-sweep
- status
- candidate
capabilities
draft launch
- accelerator backend
- nvidia
- arguments
- serve, --model, Infatoshi/Qwen3.6-35B-A3B-NVFP4-FP8, --tensor-parallel-size, 1, --host, 0.0.0.0, --port, 8000, --max-model-len, 262144
- 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 only | target |
|---|---|
| No | /root/.cache/huggingface |
engine
- name
- vllm
- version
- 0.20.2rc1.dev189+gfb1ac806c
serving
- max concurrency
- 8
- max context tokens
- 262,144
- 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
- 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
- cuda
- localmaxxing · batch size
- 1
- localmaxxing · hardware label
- RTX 5090
- localmaxxing · notes
- Automated benchmark via agent. Model: Qwen3.6-35B-A3B, Infatoshi NVFP4-FP8 (mixed-precision: NVFP4 MoE experts + FP8 attention), Docker vllm/vllm-openai:nightly, MTP spec decode.
- localmaxxing · observed command
- vllm serve /models/Qwen3.6-35B-A3B-NVFP4-FP8 --served-model-name Qwen3.6-35B-A3B --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder --max-model-len auto --limit-mm-per-prompt.video 0 --max-cudagraph-capture-size 32 --max-num-seqs 8 --max-num-batched-tokens 4096 --gpu-memory-utilization 0.940 --trust-remote-code --quantization compressed-tensors --safetensors-load-strategy prefetch --enable-chunked-prefill --enable-prefix-caching --scheduling-policy priority --language-model-only --attention-backend flash_attn --speculative-config {"method":"qwen3_next_mtp","num_speculative_tokens":4}
- localmaxxing · run id
- cmp123ars00y0lo01mj82rijd
- localmaxxing · tokenized · arguments
- vllm, serve, /models/Qwen3.6-35B-A3B-NVFP4-FP8, --served-model-name, Qwen3.6-35B-A3B, --reasoning-parser, qwen3, --enable-auto-tool-choice, --tool-call-parser, qwen3_coder, --max-model-len, auto, --limit-mm-per-prompt.video, 0, --max-cudagraph-capture-size, 32, --max-num-seqs, 8, --max-num-batched-tokens, 4096, --gpu-memory-utilization, 0.940, --trust-remote-code, --quantization, compressed-tensors, --safetensors-load-strategy, prefetch, --enable-chunked-prefill, --enable-prefix-caching, --scheduling-policy, priority, --language-model-only, --attention-backend, flash_attn, --speculative-config, {method:qwen3_next_mtp,num_speculative_tokens:4}
- 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/cmp123ars00y0lo01mj82rijd ↗ |