4 × NVIDIA GB200

Tested system

Processor: NVIDIA Grace CPU
Graphics: NVIDIA GB200
Accelerator count: 4
Nodes: 1

Recorded results

Only the same benchmark, version, workload, scenario and unit share a row. Missing results are not zero. System results include the full submitted hardware and software configuration.

MLPerf Inference / gpt-oss-120bOffline · v6.1 · Tokens/s

4 × NVIDIA GB200

61,205.60^* Tokens/s

Accuracy target / result: exact_match: 83.695
Weight precision: fp8
Division: closed
System type: datacenter
Availability: available
Cores per processor: 72
Processors per node: 2
Nodes: 1
Total accelerators: 4
Software stack: vLLM 0.24.0, Red Hat OpenShift 4.21
Operating system: Red Hat Enterprise Linux CoreOS 9.6

MLPerf Inference / gpt-oss-120bServer · v6.1 · Tokens/s

4 × NVIDIA GB200

51,095.20^* Tokens/s

Accuracy target / result: exact_match: 83.948
Weight precision: fp8
Division: closed
System type: datacenter
Availability: available
Cores per processor: 72
Processors per node: 2
Nodes: 1
Total accelerators: 4
Software stack: vLLM 0.24.0, Red Hat OpenShift 4.21
Operating system: Red Hat Enterprise Linux CoreOS 9.6

Processing & platform

Processor: NVIDIA Grace CPU
Graphics: NVIDIA GB200

Connectivity & software

Operating system: Red Hat Enterprise Linux CoreOS 9.6

System configuration

Reported host core count: 72
Processors per node: 2
Accelerator count: 4
Nodes: 1
Software stack: vLLM 0.24.0, Red Hat OpenShift 4.21
Submission division: closed

* Provisional data