Submitter Intel
Submission ID c14186d2-0438-4b2b-a204-6c66dfbd5365
Submission Status PUBLISHED
Submission Date 2026-06-26
Publication Date —
Division Standardized
Model Name GPT-OSS 120B
Division Standardized
Model Precision FP16
Model Link —
Calibration Link —
Model Notes —
Dataset Name GPT-OSS Dataset
Dataset Type performance
Average Input Tokens 5011
Dataset Link —
System TPS vs. Tokens/s per User
System TPS vs. Concurrency
Time to First Token P95 vs. Concurrency
Tokens/s per User vs. Concurrency
System Name 4x Arc(TM) Pro B70
System Availability Available
System Category Datacenter · —
System Size 4x Intel(R) Arc(TM) Pro B70 Graphics
Node Types 1
Total Node Count 1
Max Supported Concurrency 1,536
Node Type ID 0
Node Count 1
Framework vLLM 0.1.dev11435+gfa77e1a63
Serving Framework vLLM 0.1.dev11435+gfa77e1a63
Inference Backend Not detected: CUDA/ROCm/XPU runtime not found
Driver Not available
Link to Container —
Operating System linux debian ubuntu 24.04
Filesystem ext4 vfat
Other Software —
Software Notes —
Processor
Processor Model Name Intel(R) Xeon(R) 698X
Processors per Node 1
Cores Per Processor 86
VCPUs Per Processor 172
Accelerator
Accelerator Model Name Intel(R) Arc(TM) Pro B70 Graphics
Accelerators per Node 4
Memory Type GDDR6
Memory Capacity 32GiB
Accelerator Interconnect Not available
Host-Accelerator Interconnect PCIe 5.0 x16
Host / Storage
Host Memory Capacity 134G
Memory Configuration 8 x 16GB DDR5-6400
Storage Capacity 1.8 TB NVMe SSD
Storage Type NVMe SSD
Networking
Network Interfaces 3x Ethernet
Network Details Ethernet
Network Topology —
Other
Other Hardware —
Cooling —
Hardware Notes —
| Field Name | Run 1 | Run 2 | Run 3 | Run 4 | Run 5 | Run 6 | Run 7 |
|---|---|---|---|---|---|---|---|
| Run Date | 06/26/2026 | 06/26/2026 | 06/26/2026 | 06/26/2026 | 06/26/2026 | 06/26/2026 | 06/26/2026 |
| Concurrency | 16 | 32 | 40 | 64 | 128 | 256 | 384 |
| System Tokens/Second | 311.5 | 545.4 | 663.1 | 937.8 | 1,326.3 | 1,522.5 | 1,546.8 |
| Tokens/Second per User | 19.5 | 17.0 | 16.6 | 14.7 | 10.4 | 5.9 | 4.0 |
| TTFT P95 (ms) | 1210.7 | 1312.9 | 1362.5 | 1605.1 | 2172.4 | 4430.7 | 36237.6 |
| Utilization | 20.1% | 35.3% | 42.9% | 60.6% | 85.7% | 98.4% | 100.0% |
| Configuration Summary | gpt-oss-120b-benchmark · online · c16 | gpt-oss-120b-benchmark · submission · c32 | gpt-oss-120b-benchmark · submission · c40 | gpt-oss-120b-benchmark · submission · c64 | gpt-oss-120b-benchmark · submission · c128 | gpt-oss-120b-benchmark · submission · c256 | gpt-oss-120b-benchmark · submission · c384 |
| Config Type | online | submission | submission | submission | submission | submission | submission |
| Config Region | — | — | — | — | — | — | — |
| Target QPS | — | — | — | — | — | — | — |
| Result State | complete | complete | complete | complete | complete | complete | complete |
| Measured Accuracy Score | — | — | — | — | — | — | — |
| Benchmark Version | ba95c2f | 7251d70 | 7251d70 | 7251d70 | 7251d70 | 7251d70 | 7251d70 |
| Result Version | 0.1.0 | 0.1.0 | 0.1.0 | 0.1.0 | 0.1.0 | 0.1.0 | 0.1.0 |
| Archive URI | gs://mlperf-endpoints-blob/runs/0e3682e7-5b5e-4dbb-a068-e68c380b7e00/archive.tar.gz | gs://mlperf-endpoints-blob/runs/4a4bce79-4897-4223-aa6f-f166131df2c6/archive.tar.gz | gs://mlperf-endpoints-blob/runs/85a2b623-000c-4010-9884-b02f484112f4/archive.tar.gz | gs://mlperf-endpoints-blob/runs/8fa8a7df-685f-4df5-ab05-df89fbd772f0/archive.tar.gz | gs://mlperf-endpoints-blob/runs/81faee26-b00c-4e49-b44d-612dc48538dc/archive.tar.gz | gs://mlperf-endpoints-blob/runs/a30b386b-77ee-4a1f-baf6-7016f0b1ee15/archive.tar.gz | gs://mlperf-endpoints-blob/runs/bd5d352a-2ef9-4499-8373-412cd7724cd1/archive.tar.gz |
| QPS | 0.2291 | 0.3945 | 0.4831 | 0.6908 | 0.9780 | 1.1287 | 1.1512 |
| Total Output Tokens | 8,698,204 | 8,841,004 | 8,779,517 | 8,682,555 | 8,673,438 | 8,627,637 | 8,594,475 |
| Average Output Tokens | 1,359.94 | 1,382.27 | 1,372.66 | 1,357.5 | 1,356.07 | 1,348.91 | 1,343.73 |
| Run Duration (s) | 27,922.44 | 16,211.49 | 13,240.59 | 9,258.68 | 6,539.63 | 5,666.91 | 5,556.15 |
| Total Queries | 6,396 | 6,396 | 6,396 | 6,396 | 6,396 | 6,396 | 6,396 |
| Total Queries Issued | 6,396 | 6,396 | 6,396 | 6,396 | 6,396 | 6,396 | 6,396 |
| Total Queries Failed | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Time to First Token (TTFT) P95 (ms) | |||||||
| Minimum | 216.7 | 216.0 | 217.4 | 239.4 | 290.2 | 397.1 | 642.1 |
| Average | 678.8 | 748.9 | 813.5 | 999.2 | 1748.7 | 4505.6 | 20619.4 |
| P50 | 629.7 | 662.4 | 675.5 | 715.2 | 949.7 | 1482.2 | 14702.6 |
| P90 | 1041.8 | 1110.6 | 1137.1 | 1339.8 | 1766.6 | 3104.2 | 26998.5 |
| P95 | 1210.7 | 1312.9 | 1362.5 | 1605.1 | 2172.4 | 4430.7 | 36237.6 |
| P99 | 1603.5 | 1782.7 | 1878.2 | 2888.6 | 35788.4 | 108954.3 | 194077.1 |
| P999 | 5433.2 | 11943.3 | 19032.2 | 31341.5 | 68319.3 | 143341.6 | 233837.9 |
| Maximum | 8185.5 | 15214.2 | 22701.1 | 35584.5 | 72180.0 | 148949.9 | 239378.9 |
| Time Per Output Token (TPOT) (ms) | |||||||
| Minimum | 46.2 | 50.3 | 48.1 | 50.3 | 49.7 | 60.9 | 66.9 |
| Average | 50.8 | 58.0 | 59.6 | 67.2 | 94.3 | 162.4 | 230.1 |
| P50 | 50.8 | 57.9 | 59.6 | 67.3 | 94.6 | 164.3 | 235.1 |
| P90 | 51.9 | 59.5 | 61.2 | 69.3 | 97.9 | 168.4 | 238.8 |
| P95 | 52.2 | 60.1 | 61.8 | 69.9 | 98.9 | 169.5 | 240.0 |
| P99 | 52.8 | 61.1 | 62.9 | 71.1 | 101.0 | 172.2 | 242.3 |
| P999 | 53.6 | 62.7 | 65.0 | 72.9 | 104.1 | 177.2 | 245.0 |
| Maximum | 54.8 | 114.4 | 113.1 | 74.1 | 108.5 | 190.3 | 250.0 |
| Query Latency (ms) | |||||||
| Minimum | 30577.8 | 1383.0 | 1063.2 | 34035.6 | 47246.1 | 71126.3 | 96318.8 |
| Average | 69731.1 | 80857.2 | 82551.9 | 92084.2 | 129490.3 | 223274.1 | 329335.2 |
| P50 | 68583.0 | 78453.3 | 80475.9 | 90583.4 | 127444.4 | 219880.2 | 324416.1 |
| P90 | 86677.9 | 100534.8 | 103337.2 | 115851.3 | 162988.7 | 283233.2 | 418366.6 |
| P95 | 94001.0 | 108760.4 | 110007.0 | 124458.1 | 175818.0 | 306836.1 | 451690.7 |
| P99 | 108874.9 | 130471.9 | 127699.9 | 141476.7 | 204148.0 | 360063.9 | 539161.6 |
| P999 | 128584.3 | 589166.0 | 607952.9 | 173235.0 | 240468.8 | 411631.2 | 613361.8 |
| Maximum | 520748.0 | 593328.8 | 617914.4 | 192981.5 | 323771.5 | 477606.0 | 709762.9 |