CoreWeave · CoreWeave GB200 NVL72 (64x GB200-186GB_aarch64, TensorRT)
GPT-OSS-120B (MLPerf reference) · gpt-oss-120b · Offline · run 01a00b49-c254…
valid · Reported evidence
Measurements
output_token_throughput_tps · reported774,539 tokens_s
GPT-OSS-120B (MLPerf reference) · gpt-oss-120b · Offline · offline · 4× NVIDIA GB200 × 16 nodes · mlperf_closed_99 · observed 2026-03-26
Identity
modelGPT-OSS-120B (MLPerf reference)
stackTensorRT 10.14, CUDA 13.1
workloadgpt-oss-120b · Offline
topology4× NVIDIA GB200 × 16 nodes
harnessmlperf_loadgen v6.0
quality policymlperf_closed_99
comparison keysha256:9fd27c4b8f6938d4…
sha256:37290147d02fd268b0d59ec…
Launch configuration
dtypefp4
notesGB200 NVL72
systemCoreWeave GB200 NVL72 (64x GB200-186GB_aarch64, TensorRT)
softwareTensorRT 10.14, CUDA 13.1
availabilityavailable
Workload
namegpt-oss-120b · Offline
streamingno
load generationoffline
Canonical manifest
Show manifestHide manifest
{
"kind": "ServingRun",
"spec": {
"status": "valid",
"harness": {
"name": "mlperf_loadgen",
"version": "v6.0"
},
"topology": {
"nodeCount": 16,
"acceleratorModel": "NVIDIA GB200",
"acceleratorVendor": "nvidia",
"acceleratorsPerNode": 4
},
"observedAt": "2026-03-26T19:39:54Z",
"modelDigest": "sha256:25d396066173182e38fe499acf87750416b561f71db6feef2907188a1d59dff7",
"measurements": [
{
"unit": "tokens_s",
"value": 774539,
"metric": "output_token_throughput_tps",
"statistic": "reported"
}
],
"sourceNative": {
"source": "mlperf-inference",
"metrics": {
"performance_units": "Tokens/s",
"performance_result": 774539
},
"category": "datacenter",
"division": "closed",
"scenario": "offline",
"benchmark": "gpt-oss-120b",
"externalId": "v6.0/6.0-0015"
},
"qualityPolicy": "mlperf_closed_99",
"workloadDigest": "sha256:51bf73d7ec6f9d34eac7dec515275ae3acd0890eeab2278cd6c99b7a889352ee",
"configurationDigest": "sha256:ed7d132384d0c5e3b48bc9251b10de4dd43b46ee2f5c510b1eb1056641c70980"
},
"metadata": {
"name": "mlperf-v6-0-6-0-0015",
"title": "CoreWeave · gpt-oss-120b · Offline",
"sourceRefs": [
{
"url": "https://github.com/mlcommons/submissions_inference_v6.0/tree/main/closed/CoreWeave/results/GB200-NVL72_GB200-186GB_aarch64x64_TRT"
}
]
},
"apiVersion": "kernelindex.dev/v1alpha1"
}Cite this result (permalink, digest, access date)
Report an issue with this serving run
MLPerf Inference · closed/datacenter · entry v6.0/6.0-0015 · shown unmodified as publishedAll serving results →JSON