Local AI Registry

Recipe

qwen3-6-35b-a3b-fp8-online-fp8-kv-intel-arc-pro-b70-32gb-vllm-tp2

qwen3-6-35b-a3b-fp8-online-fp8-kv-intel-arc-pro-b70-32gb-vllm-tp2
JSON APIFind compatible recipes

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

Record

Status
candidate
Source
localmaxxing
Engine
vllm
Engine version
vllm-main d272418f4 (XPU graph + PatchB)
Graph
full-decode-only
Accelerators
2
Tensor parallel
2
Context tokens
65,536
Max concurrency
8
chat
unknown
reasoning
unknown
tools
unknown
vision
unknown

Hugging Face model card

Identity

https://huggingface.co/Qwen/Qwen3.6-35B-A3B
Repository
Qwen/Qwen3.6-35B-A3B
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

Source
https://www.localmaxxing.com/en/runs/cmqsyq9lt011rqr010otg1494

Observed source tokens

Mechanical split of the source command. Unverified against the engine CLI. The registry does not offer Run for this recipe.

  1. vllm
  2. serve
  3. qwen36-35b-a3b-bf16-hf
  4. --quantization
  5. fp8
  6. --kv-cache-dtype
  7. fp8
  8. --enable-prefix-caching
  9. --mamba-cache-mode
  10. align
  11. --tensor-parallel-size
  12. 2
  13. --dtype
  14. float16
  15. --max-model-len
  16. 65536
  17. --max-num-seqs
  18. 8
  19. --gpu-memory-utilization
  20. 0.90
  21. --max-num-batched-tokens
  22. 8192
  23. --trust-remote-code
  24. --enable-auto-tool-choice
  25. --tool-call-parser
  26. qwen3_coder
  27. --compilation-config
  28. {"cudagraph_mode":"FULL_DECODE_ONLY","cudagraph_capture_sizes":[1,2,4,8]}
FlagValue
--quantizationfp8
--kv-cache-dtypefp8
--mamba-cache-modealign
--tensor-parallel-size2
--dtypefloat16
--max-model-len65536
--max-num-seqs8
--gpu-memory-utilization0.90
--max-num-batched-tokens8192
--tool-call-parserqwen3_coder
--compilation-config{"cudagraph_mode":"FULL_DECODE_ONLY","cudagraph_capture_sizes":[1,2,4,8]}

Environment

VariableValue
VLLM_XPU_ENABLE_XPU_GRAPH1
ZE_AFFINITY_MASK0,2

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatusSweep
—65,536—110.7—observedqwen3-6-35b-a3b-fp8-online-fp8-kv-intel-arc-pro-b70-32gb-vllm-tp2-sweep

Remaining fields

Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.

hardware count
2
hardware id
intel-arc-pro-b70-32gb
id
qwen3-6-35b-a3b-fp8-online-fp8-kv-intel-arc-pro-b70-32gb-vllm-tp2
model instance id
qwen-qwen3-6-35b-a3b--fp8-online-fp8-kv
recipe source
localmaxxing
schema version
local-ai-registry/v1
speed sweep ids
qwen3-6-35b-a3b-fp8-online-fp8-kv-intel-arc-pro-b70-32gb-vllm-tp2-sweep
status
candidate

capabilities

engine

graph mode
full-decode-only
name
vllm
version
vllm-main d272418f4 (XPU graph + PatchB)

serving

max concurrency
8
max context tokens
65,536
tensor parallel
2
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
2026-09-01T01:32:29Z
engine.graph mode · reason explicit-cudagraph-mode-in-observed-command
engine.graph mode · 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
2026-09-01T10:20:39Z
serving.max concurrency · reason explicit-source-server-capacity
serving.max concurrency · state known
serving.max context tokens · provenance · captured at
2026-09-01T10:20:39Z
serving.max context tokens · reason explicit-source-context-limit
serving.max context tokens · state known

metadata

localmaxxing · backend
xpu
localmaxxing · batch size
1
localmaxxing · hardware label
Intel Arc Pro B70
localmaxxing · notes
b1 DECODE = 110.73 t/s (corrected after-first-token, mean-of-5, temp 0, deterministic 110.71-110.76). Qwen3.6-35B-A3B online-FP8 + FP8-KV on 2x Intel Arc Pro B70 (TP=2) via vLLM-XPU graph-mode (FULL_DECODE_ONLY, vllm-main d272418f4 + PatchB kernel). Cold prefill ~6000 tok/s; align prefix-cache + qwen3_coder tool-calling. Higher-batch aggregate (1 TP=2 instance): b8=592, b16=598, b24=603 tok/s. Stability: ~20h continuous SWE-bench-Verified-500 agentic load, zero corruption/repetition. (A native-BF16 TP=4 config reaches 120.76 b1 -- separate entry.) HP B70 variant excluded; no power-limit changes.
localmaxxing · observed command
ZE_AFFINITY_MASK=0,2 VLLM_XPU_ENABLE_XPU_GRAPH=1 vllm serve qwen36-35b-a3b-bf16-hf --quantization fp8 --kv-cache-dtype fp8 --enable-prefix-caching --mamba-cache-mode align --tensor-parallel-size 2 --dtype float16 --max-model-len 65536 --max-num-seqs 8 --gpu-memory-utilization 0.90 --max-num-batched-tokens 8192 --trust-remote-code --enable-auto-tool-choice --tool-call-parser qwen3_coder --compilation-config '{"cudagraph_mode":"FULL_DECODE_ONLY","cudagraph_capture_sizes":[1,2,4,8]}'
localmaxxing · run id
cmqsyq9lt011rqr010otg1494
localmaxxing · tokenized · arguments
vllm, serve, qwen36-35b-a3b-bf16-hf, --quantization, fp8, --kv-cache-dtype, fp8, --enable-prefix-caching, --mamba-cache-mode, align, --tensor-parallel-size, 2, --dtype, float16, --max-model-len, 65536, --max-num-seqs, 8, --gpu-memory-utilization, 0.90, --max-num-batched-tokens, 8192, --trust-remote-code, --enable-auto-tool-choice, --tool-call-parser, qwen3_coder, --compilation-config, {"cudagraph_mode":"FULL_DECODE_ONLY","cudagraph_capture_sizes":[1,2,4,8]}
localmaxxing · tokenized · environment · VLLM XPU ENABLE XPU GRAPH
1
localmaxxing · tokenized · environment · ZE AFFINITY MASK
0,2
localmaxxing · tokenized · fidelity
faithful

provenance

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

sources

captured atkindurl
2026-08-30T09:10:02Znormalized-recipewww.localmaxxing.com/en/runs/cmqsyq9lt011rqr010otg1494 ↗