vLLM/recipes/models/Qwen/QwQ-32B.yaml

meta:
  title: "QwQ-32B"
  slug: "qwq-32b"
  provider: "Qwen"
  description: "Qwen's 32B dense reasoning model, validated for Intel Xeon 6 CPU execution."
  date_added: 2026-08-17
  date_updated: 2026-09-12
  tasks:
    - text
  performance_headline: "Validated for Intel Xeon 6 CPU execution"
  related_recipes:
    - "Qwen/Qwen3-32B"
  hardware:
    xeon6: verified

model:
  model_id: "Qwen/QwQ-32B"
  min_vllm_version: "0.7.3"
  architecture: dense
  parameter_count: "32B"
  active_parameters: "32B"
  context_length: 40960
  base_args: []
  base_env: {}

features:
  tool_calling:
    description: "Enable automatic tool choice with the Hermes parser"
    args:
      - "--enable-auto-tool-choice"
      - "--tool-call-parser"
      - "hermes"
  reasoning:
    description: "Extract QwQ reasoning content with the DeepSeek-R1 reasoning parser"
    args:
      - "--reasoning-parser"
      - "deepseek_r1"
  long_context:
    description: "YaRN RoPE scaling to the full 131,072-token context, which the model card requires for prompts over 8,192 tokens. Static YaRN holds the scaling factor constant regardless of input length, so leave this off unless you actually serve long prompts."
    args:
      - "--hf-overrides"
      - '{"max_position_embeddings":131072,"rope_parameters":{"rope_type":"yarn","factor":4.0,"original_max_position_embeddings":32768}}'
      - "--max-model-len"
      - "131072"

opt_in_features:
  - long_context

variants:
  default:
    precision: bf16
    vram_minimum_gb: 77
    description: "Full precision BF16"
    supported_hardware:
      - xeon6
  w8a8_int8:
    model_id: "RedHatAI/QwQ-32B-quantized.w8a8"
    precision: int8
    vram_minimum_gb: 39
    description: "Red Hat AI INT8 W8A8 compressed-tensors checkpoint"
    supported_hardware:
      - xeon6
  awq:
    model_id: "Qwen/QwQ-32B-AWQ"
    precision: int4
    vram_minimum_gb: 20
    description: "Qwen official AWQ 4-bit checkpoint"
    supported_hardware:
      - xeon6
    extra_args:
      - "--quantization"
      - "awq"

compatible_strategies:
  - single_node_tp
  - multi_node_tp

kv_offload_support:
  offloading_cpu: verified
  offloading_fs: verified

hardware_overrides: {}

strategy_overrides: {}

guide: |
  ## Overview

  QwQ-32B is Qwen's 32B dense reasoning model. The CPU model validation table
  marks it supported on Intel Xeon 6.

  ## Prerequisites

  - Hardware: Intel Xeon 6 CPUs
  - vLLM CPU >= 0.8.5 for this validated Xeon 6 configuration

  ### pip (Intel Xeon 6 CPUs)

  For Intel and AMD x86 CPUs, follow the
  [CPU pre-built wheels](https://docs.vllm.ai/en/latest/getting_started/installation/cpu/#pre-built-wheels)
  installation instructions.

  ### Docker (Intel Xeon 6 CPUs)

  ```bash
  docker pull vllm/vllm-openai-cpu:latest-x86_64
  ```

  ## Intel Xeon 6

  ```bash
  vllm serve Qwen/QwQ-32B \
    --tensor-parallel-size 1 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --reasoning-parser deepseek_r1
  ```

  Docker (the image entrypoint is `vllm serve`):

  ```bash
  docker run \
    --privileged --ipc=host -p 8000:8000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    vllm/vllm-openai-cpu:latest-x86_64 Qwen/QwQ-32B \
    --tensor-parallel-size 1 \
    --enable-auto-tool-choice \
    --tool-call-parser hermes \
    --reasoning-parser deepseek_r1
  ```

  W8A8 INT8:

  ```bash
  vllm serve RedHatAI/QwQ-32B-quantized.w8a8
  ```

  The example uses TP=1. Adjust tensor parallelism based on system topology;
  topology-specific CPU binding is intentionally not hard-coded.

  ## Runtime and Platform Tuning

  The following settings are intentionally not prescribed as portable CPU model
  defaults because they depend on the workload, hardware, or runtime environment:

  - `--max-num-batched-tokens`: scheduler/throughput tuning.
  - `--max-num-seqs`: concurrency and scheduler-capacity tuning.
  - `--gpu-memory-utilization`: platform memory-budget tuning.
  - `--no-enable-prefix-caching`: workload/benchmark cache-policy tuning.
  - `VLLM_ENGINE_ITERATION_TIMEOUT_S`: operational runtime timeout.

  These settings may still appear in validated hardware-specific overrides. Tune
  them at deployment time based on platform resources, workload shape, and
  latency/throughput goals.

  ## References

  - [Model card](https://huggingface.co/Qwen/QwQ-32B)
  - [W8A8 variant](https://huggingface.co/RedHatAI/QwQ-32B-quantized.w8a8)
  - [vLLM reasoning outputs](https://docs.vllm.ai/en/latest/features/reasoning_outputs.html)
  - [vLLM tool calling](https://docs.vllm.ai/en/latest/features/tool_calling.html)