1 GPU on the selected test

Qwen3.8-27B-NVFP4

GPUs are compared only on one exact test of this model. Change the test to see other conditions.

Run this model

Choose your operating system and enter a model repository. Launch support is checked separately from performance results.

Test conditions

Every GPU below ran exactly these conditions.

Task: Text generation · Text → Text
Workload: 128K → 2K tokens · No cache reuse
Concurrency: 1 request
Format · Engine: NVFP4 · vllm · vLLM 0.28.0
Unit: tok/s

Other test conditions

Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.

Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.

Protocol details

wbench-search-http-v1 · migration-analysis-v1:sha256=68590c248e43f32c04dfd8062464fd1e4856917795fe307225e5c6112a54fe94:cache=isolated:repeats=3 · 131072in-2048out

GPUs on this test

Only one GPU configuration has matching measurements; a cross-hardware comparison is not yet available.

RTX PRO 6000 Blackwell Server Edition · 96 GB
Mean throughput
31.5tok/s
Observed GPU occupancy
87.4GiB
First token
24.9 s
Reports combined
1
Runs, parameters and setup ↗

Sortable table

Hardware ranking · 1 rows
Hardware	Mean throughput (tok/s)	P95 (s)	TTFT (ms)	Observed occupancy (GiB)	GPU power (W)	Throughput / W	VRAM per GPU (GB)	Reports combined
#1RTX PRO 6000 Blackwell Server Edition · 96 GBRuns, parameters and setup ↗	31.54	64.932	24,870	87.37	402.69	0.08	96	1

Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.

Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.

How results are grouped

Results merge successful runs of the same model, precision and workload. CPU thread counts and CUDA wheel builds may differ; actual PyTorch builds are shown. Other recorded protocol settings remain separate.

Updated 11/5/2026, 9:56:52 PM