Menu
Back previous/next report: k j

NVIDIA NVIDIA GB300 NVL72

Submitted by NVIDIAon 2026-06-25

Submission Info
Submitter NVIDIA
Submission ID 27168b87-e0f9-4488-ae93-88a45c1c81c9
Submission Status PUBLISHED
Submission Date 2026-06-25
Publication Date
Division Standardized
Model
Model Name DeepSeek R1
Division Standardized
Model Precision
Model Link
Calibration Link
Model Notes
Datasets
Dataset Name dsr1-perf
Dataset Type performance
Average Input Tokens
Dataset Link

System TPS vs. Tokens/s per User

0200k400k600kSystem TPS0100200300Tokens/s per User

System TPS vs. Concurrency

0200k400k600kSystem TPS010,00020,00030,000# of queries

Time to First Token P95 vs. Concurrency

010203040TTFT P95 (s)010,00020,00030,000# of queries

Tokens/s per User vs. Concurrency

0100200300Tokens/s per User010,00020,00030,000# of queries
SUT Summary
NVIDIA
System Name NVIDIA GB300 NVL72
System Availability Available
System Category Datacenter · GB300 NVL72 rack, disaggregated TRT-LLM, NVFP4 weights
System Size
Node Types 1
Total Node Count 18
Max Supported Concurrency 28,080
Node Type ID 0
Node Count 18
Software
Framework TensorRT-LLM
Serving Framework TensorRT-LLM
Inference Backend TensorRT-LLM
Driver
Link to Container
Operating System unspecified
Filesystem
Other Software
Software Notes
Hardware

Processor

Processor Model Name NVIDIA Grace (ARM Neoverse V2)
Processors per Node 1
Cores Per Processor 72
VCPUs Per Processor

Accelerator

Accelerator Model Name NVIDIA GB300
Accelerators per Node 4
Memory Type
Memory Capacity 288 GB HBM3e
Accelerator Interconnect NVLink 5 + NVSwitch (NVL72)
Host-Accelerator Interconnect NVLink-C2C (Grace-Blackwell)

Host / Storage

Host Memory Capacity unspecified
Memory Configuration
Storage Capacity unspecified
Storage Type unspecified

Networking

Network Interfaces
Network Details NVLink + InfiniBand
Network Topology NVL72 single-rack NVLink domain

Other

Other Hardware
Cooling Liquid-cooled
Hardware Notes
Run Data
Field NameRun 1Run 2Run 3Run 4Run 5Run 6Run 7Run 8Run 9Run 10Run 11Run 12Run 13Run 14
Run Date06/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/202606/25/2026
Concurrency12448641282565121,0242,0484,0968,19216,38428,080
System Tokens/Second338.8640.91,225.110,556.514,161.027,256.645,293.781,627.8130,376.4198,881.9306,707.1395,958.1469,668.6546,980.1
Tokens/Second per User338.8320.5306.3219.9221.3212.9176.9159.4127.397.174.948.328.719.5
TTFT P95 (ms)1193.01065.0317.5467.2451.2521.62679.3520.1572.11306.44550.57084.919877.036763.2
Utilization0.1%0.1%0.2%1.9%2.6%5.0%8.3%14.9%23.8%36.4%56.1%72.4%85.9%100.0%
Configuration Summarydsr1-gb300x72-pareto-nvl72-c1 · submissiondsr1-gb300x72-pareto-nvl72-c2 · submissiondsr1-gb300x72-pareto-nvl72-c4 · submissiondsr1-gb300x72-pareto-nvl72-c48 · submissiondsr1-gb300x72-pareto-nvl72-c64 · submissiondsr1-gb300x72-pareto-nvl72-c128 · submissiondsr1-gb300x72-pareto-nvl72-c256 · submissiondsr1-gb300x72-pareto-nvl72-c512 · submissiondsr1-gb300x72-pareto-nvl72-c1024 · submissiondsr1-gb300x72-pareto-nvl72-c2048 · submissiondsr1-gb300x72-pareto-nvl72-c4096 · submissiondsr1-gb300x72-pareto-nvl72-c8192 · submissiondsr1-gb300x72-pareto-nvl72-c16384 · submissiondsr1-gb300x72-pareto-nvl72-c28080-numa · submission
Config Typesubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmissionsubmission
Config Region
Target QPS
Result Statecompletecompletecompletecompletecompletecompletecompletecompletecompletecompletecompletecompletecompletecomplete
Measured Accuracy Score
Benchmark Version05ce40005ce40005ce400ced1dcded0c53d9019316cd477d4429c5b89cbfefeaf6bf98cd477d40470dab7265704e072ac2
Result Version0.1.00.1.00.1.00.1.00.1.00.1.00.1.00.1.00.1.00.1.00.1.00.1.00.1.00.1.0
Archive URIgs://mlperf-endpoints-blob/runs/49655a2c-3d3d-45f4-916e-54c439080875/archive.tar.gzgs://mlperf-endpoints-blob/runs/00747f95-5433-4718-99a6-17c91e6000bb/archive.tar.gzgs://mlperf-endpoints-blob/runs/cbe204a9-7436-4ff0-85b0-a7b37b4b785c/archive.tar.gzgs://mlperf-endpoints-blob/runs/6ae0b5b8-33d9-4640-af10-5adfa595b66f/archive.tar.gzgs://mlperf-endpoints-blob/runs/4f76a1f0-e4a7-4fc1-ab15-f767674462b2/archive.tar.gzgs://mlperf-endpoints-blob/runs/1d17bdfe-f13c-44c3-beda-3041f3cfd2ae/archive.tar.gzgs://mlperf-endpoints-blob/runs/a913ac1d-7f38-453a-ba89-e0cbd1ab5250/archive.tar.gzgs://mlperf-endpoints-blob/runs/aa5cc47c-c121-41c2-99d8-ccd4bbd607b9/archive.tar.gzgs://mlperf-endpoints-blob/runs/a9c5e6a2-71f1-4d0d-8302-f1cc94a43ab4/archive.tar.gzgs://mlperf-endpoints-blob/runs/9784c932-acdd-46de-9096-fbb6a1e4a723/archive.tar.gzgs://mlperf-endpoints-blob/runs/523ba87c-7040-4147-bb76-360ae7979a56/archive.tar.gzgs://mlperf-endpoints-blob/runs/ec5cf744-36e8-4e9d-8356-618204ec2339/archive.tar.gzgs://mlperf-endpoints-blob/runs/164c6f8d-96d1-4aab-b7e1-5fbcf7d1ff40/archive.tar.gzgs://mlperf-endpoints-blob/runs/68478c69-8d3c-4be8-82d9-8f79a3673c63/archive.tar.gz
QPS0.08870.15680.33052.82983.77597.274311.937621.703634.766552.957881.3117105.4162125.2644145.1861
Total Output Tokens954,7452,043,7243,706,83324,553,92438,253,79382,208,88283,244,526280,557,875394,925,643725,078,1311,075,847,7201,318,555,9251,349,790,3442,479,730,589
Average Output Tokens3,818.984,087.453,706.833,730.473,750.373,746.993,794.193,761.033,750.053,755.483,771.993,756.173,749.423,767.44
Run Duration (s)2,817.613,188.723,025.622,325.962,701.363,016.111,837.883,437.043,029.123,645.773,507.743,330.042,873.924,533.49
Total Queries2505001,0006,58210,20021,94021,94074,596105,312193,072285,220351,040360,000658,200
Total Queries Issued2505001,0006,58210,20021,94021,94074,596105,312193,072285,220351,040360,000658,200
Total Queries Failed00000000000000
Time to First Token (TTFT) P95 (ms)
Minimum143.7140.7201.4164.1165.0168.8142.8179.5157.9215.6174.1180.1217.2176.2
Average325.9276.7274.5312.6307.6368.6776.8397.3388.9608.01109.71857.64698.16033.8
P50229.9233.3239.5268.1266.3342.7329.7391.8329.2483.6430.9454.4535.7356.4
P901123.1251.8259.1432.8423.9492.72051.5491.5541.0989.22851.95612.915454.015410.8
P951193.01065.0317.5467.2451.2521.62679.3520.1572.11306.44550.57084.919877.036763.2
P991312.31269.91284.1534.2522.0579.43902.9599.4726.42272.38526.018516.046295.082744.4
P9991331.21340.41873.21536.61464.41959.15126.22912.54622.46870.215467.227366.355336.399115.9
Maximum1332.31395.61875.51537.31465.21994.55791.32984.84636.57897.716710.528908.957636.7101763.9
Time Per Output Token (TPOT) (ms)
Minimum0.000.100.672.942.933.101.734.094.724.542.785.135.4512.6
Average2.863.093.304.494.494.635.356.217.7410.112.919.231.845.5
P502.903.083.264.424.434.575.306.147.6510.012.819.132.147.3
P903.223.543.805.004.995.135.926.878.5611.314.421.536.547.9
P953.293.624.025.265.165.286.037.048.7611.514.722.037.348.1
P993.393.804.255.905.695.836.247.509.2311.915.222.838.848.5
P9993.413.874.396.616.136.446.458.4410.712.415.824.040.649.1
Maximum3.463.904.418.206.466.936.859.8212.413.517.3431.844.051.5
Query Latency (ms)
Minimum1047.51073.3875.71056.9846.91028.11214.21173.71446.51795.72097.42405.43453.96473.4
Average11270.312666.912075.116637.716696.317258.420562.923107.928604.737600.848620.972122.0120823.5177589.8
P506101.37498.07388.89763.99667.09991.912080.713400.716651.721644.828269.441994.971387.1103139.0
P9026705.534241.329495.540164.041694.642185.749759.956725.070369.892578.4118723.6176638.6292831.2435537.8
P9537603.645757.641506.158236.558608.260123.571469.280895.899463.5131976.2169503.8250693.8411163.5613731.5
P9954126.958665.159404.181313.882159.684936.999759.1114054.1141703.1186543.1240107.9356590.0594623.3937055.8
P99957437.562077.863359.193735.690663.493502.3107668.1124391.9154678.5203888.7262579.8391337.1663067.2955828.1
Maximum57773.863195.963650.6111624.6104139.5110061.7122968.0146297.9173809.6257970.0346067.0502808.5833431.1971749.4