CoreWeave · (8x GB300-288GB_aarch64, TensorRT)
DeepSeek-R1 (MLPerf reference) · deepseek-r1 · Server · run 01a00b49-a8ab…
valid · Reported evidence
Measurements
output_token_throughput_tps · reported22,545.1 tokens_s
DeepSeek-R1 (MLPerf reference) · deepseek-r1 · Server · server · 4× NVIDIA GB300 × 2 nodes · mlperf_closed_99 · observed 2025-09-04
Availability: preview.
Identity
modelDeepSeek-R1 (MLPerf reference)
stackTensorRT 10.13, CUDA 13.0
workloaddeepseek-r1 · Server
topology4× NVIDIA GB300 × 2 nodes
harnessmlperf_loadgen v5.1
quality policymlperf_closed_99
comparison keysha256:ddd1a914cdd2321c…
sha256:634cfecedf8cc55f6af5d6b…
Launch configuration
dtypefp4
notesNVIDIA GB300 NVL8
system (8x GB300-288GB_aarch64, TensorRT)
softwareTensorRT 10.13, CUDA 13.0
availabilitypreview
Workload
namedeepseek-r1 · Server
streamingyes
load generationopen_loop
Canonical manifest
Show manifestHide manifest
{
"kind": "ServingRun",
"spec": {
"status": "valid",
"caveats": [
"Availability: preview"
],
"harness": {
"name": "mlperf_loadgen",
"version": "v5.1"
},
"topology": {
"nodeCount": 2,
"acceleratorModel": "NVIDIA GB300",
"acceleratorVendor": "nvidia",
"acceleratorsPerNode": 4
},
"observedAt": "2025-09-04T19:30:14Z",
"modelDigest": "sha256:9c4d3f416e476a367a8351b29d53c726179e98b9f039335873b937afb40e7c29",
"measurements": [
{
"unit": "tokens_s",
"value": 22545.1,
"metric": "output_token_throughput_tps",
"statistic": "reported"
}
],
"sourceNative": {
"source": "mlperf-inference",
"metrics": {
"performance_units": "Tokens/s",
"performance_result": 22545.1
},
"category": "datacenter",
"division": "closed",
"scenario": "server",
"benchmark": "deepseek-r1",
"externalId": "v5.1/5.1-0097"
},
"qualityPolicy": "mlperf_closed_99",
"workloadDigest": "sha256:51ac3e54124be56b08b48d2e88fb8878fda1ff10c64431a52e66b4530fc4d7d9",
"configurationDigest": "sha256:1493260efff84ff74e3e132c1617a6a32929953e882c900e2fbebc73d9c4d50d"
},
"metadata": {
"name": "mlperf-v5-1-5-1-0097",
"title": "CoreWeave · deepseek-r1 · Server",
"sourceRefs": [
{
"url": "https://github.com/mlcommons/inference_results_v5.1/tree/main/closed/CoreWeave/results/GB300-NVL8-CoreWeave_TRT"
}
]
},
"apiVersion": "kernelindex.dev/v1alpha1"
}Cite this result (permalink, digest, access date)
Report an issue with this serving run
MLPerf Inference · closed/datacenter · entry v5.1/5.1-0097 · shown unmodified as publishedAll serving results →JSON