Qwen3.8-27B
serving recipe.

Floors are the minimum the vLLM recipe publishes across the whole replica, from vLLM's own runs. The control plane compares them with its own sizing from the model config.

• Tensor ParallelFits inside one node's interconnect. single_node_tp

• GB300 NVL4 · 4 GPUs
• RTX 5090 · 1 GPU
• RTX 5090 · 2 GPUs
• Ascend 950PR · 1 GPU

Wingsmith runs models through NVIDIA Dynamo, so it serves on NVIDIA GPUs. Other vendors are listed as the recipe states them.

vllm serve Qwen/Qwen3.8-27B \
  --enable-auto-tool-choice \
  --tool-call-parser qwen3_coder \
  --reasoning-parser qwen3 \
  --mm-encoder-tp-mode data

Features

• Tool callingOn by defaultEnable automatic tool choice. The chat template emits tool calls as <tool_call><function=…><parameter=…> XML, which is the qwen3_coder format — not JSON.
• ReasoningOn by defaultParse <think> blocks separately from the final answer. The template opens every assistant turn with <think>, so without this the reasoning lands in message.content.
• Speculative decodingOpt-inSpeculative decoding — pick MTP (in-checkpoint) or DFlash2 (external draft). Modes: MTP, MTP (Ascend), DFlash2.
• Text-only modeOpt-inSkip loading the vision encoder for text-only workloads — frees VRAM for KV cache. Mutually exclusive with encoder_parallel.
• Encoder data parallelOn by defaultRun the vision encoder in data-parallel mode — avoids TP comm overhead on the small encoder. Mutually exclusive with text_only.

Dependencies

• Matches the transformers version config.json was written by (5.8.0), for the Qwen3-VL processor classes. vLLM parses the config with its own Qwen3_5Config.

1. Register the weights

   In the console, or through the API. Resolving a slug proposes an artifact and writes nothing.

   curl "http://localhost:8000/v1/artifacts/resolve?source=hf:Qwen/Qwen3.8-27B"

2. Read the draft

   The draft carries this recipe as upstream evidence: the tested flags, the precision and memory floor, the parallelism layouts, and the hardware vLLM verified, with links back to the recipe. Nothing from it is applied without you, and the draft warns when the plane's own sizing and the published floor differ by a quarter or more.

3. Create the workload

   Choose the engine, device shape and scale. The plane picks the clusters, and each cluster's operator renders the deployment as a Dynamo graph.

4. Give it an endpoint

   One model name for clients, routed by the gateway to deployments on every cluster serving it.

Run the control plane locally →

• How much GPU memory does Qwen3.8-27B need?The vLLM recipe publishes a minimum of 24 GB across the whole replica for its INT4 variant; BF16 67 GB, FP8 38 GB, NVFP4 (W4A4) 32 GB, W8A8 (Ascend) 39 GB. That is vLLM's own measurement, not a per-GPU figure.
• Which parallelism layouts does the Qwen3.8-27B recipe support?Tensor Parallel.
• Can Wingsmith serve Qwen3.8-27B?Wingsmith serves models through NVIDIA Dynamo on clusters with NVIDIA GPUs. Register Qwen/Qwen3.8-27B and the plane attaches this vLLM recipe to the draft as evidence: its tested flags, precision and memory floor, for you to review rather than applied on your behalf.

From the vLLM recipe for Qwen/Qwen3.8-27B ↗ (snapshot at 4bcc189), Apache 2.0.