GLM-5.3-Flash
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
• Tensor + Expert ParallelFits inside one node's interconnect. single_node_tep
• Data + Expert ParallelFits inside one node's interconnect. single_node_dep
• Multi-Node Tensor ParallelSpans nodes. multi_node_tp
• Multi-Node TP + Pipeline ParallelSpans nodes. multi_node_tp_pp
• Multi-Node Tensor + Expert ParallelSpans nodes. multi_node_tep
• Multi-Node Data + Expert ParallelSpans nodes. multi_node_dep
• Prefill/Decode DisaggregationSpans nodes. pd_cluster

• B200 · 8 GPUs
• GB200 NVL4 · 4 GPUs
• H100 · 8 GPUs
• MI355X · 8 GPUs

Marked unsupported: MI300X · 8 GPUs, MI325X · 8 GPUs.

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

vllm serve zai-org/GLM-5.3-Flash \
  --tool-call-parser glm47 \
  --enable-auto-tool-choice \
  --reasoning-parser glm45

Features

• Tool callingOn by defaultGLM tool-call parsing with automatic tool choice
• ReasoningOn by defaultExtract reasoning content with the GLM reasoning parser
• Speculative decodingOpt-inUse the checkpoint's single MTP layer to draft one token

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:zai-org/GLM-5.3-Flash"

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. The one recipe default it applies, --max-num-seqs, is shown above.

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 GLM-5.3-Flash need?The vLLM recipe publishes a minimum of 229 GB across the whole replica for its NVFP4 variant; FP8 386 GB, BF16 772 GB. That is vLLM's own measurement, not a per-GPU figure.
• Which parallelism layouts does the GLM-5.3-Flash recipe support?Tensor Parallel, Tensor + Expert Parallel, Data + Expert Parallel, Multi-Node Tensor Parallel, Multi-Node TP + Pipeline Parallel, Multi-Node Tensor + Expert Parallel, Multi-Node Data + Expert Parallel and Prefill/Decode Disaggregation.
• Can Wingsmith serve GLM-5.3-Flash?Wingsmith serves models through NVIDIA Dynamo on clusters with NVIDIA GPUs. Register zai-org/GLM-5.3-Flash 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.
• Does NVIDIA publish a Dynamo recipe for GLM-5.3-Flash?Yes: 6 deployment entries for vLLM, aggregated and disaggregated. When a workload is placed, the plane defaults --max-num-seqs to the value the aggregated recipe publishes for that card, and names it on the draft.

From the vLLM recipe for zai-org/GLM-5.3-Flash ↗ (snapshot at 4bcc189) and NVIDIA Dynamo's recipes at 165d854, Apache 2.0.