Inference Chip Index

← All slices

gpt-oss-120b · Offline

gpt-oss-120b / Offline / Tokens/s

v6.0|closed|gpt-oss-120b|Offline|Tokens/s|official-system

Metric: Performance_Result (Tokens/s) · winning: higher · 28 ranking-eligible rows · vendors 3 · families 7 · never mixed with other slices

All systems (official)
RankPerformanceUnitSystemSubmitterAcceleratorSource
11,046,150.0000Tokens/sNebius GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)NebiusNVIDIA GB300source
21,042,980.0000Tokens/sNVIDIA GB300 NVL72 (72x GB300-288GB_aarch64, TensorRT)NVIDIANVIDIA GB300source
31,031,070.0000Tokens/sAMD MI355X Custom ClusterAMDAMD Instinct MI355X 288GB HBM3e (x94)source
4910,468.0000Tokens/sCoreWeave GB300 NVL72 (64x GB300-288GB_aarch64, TensorRT)CoreWeaveNVIDIA GB300source
5879,542.0000Tokens/sNVIDIA GB200 NVL72 (72x GB200-186GB_aarch64, TensorRT)NVIDIANVIDIA GB200source
6774,539.0000Tokens/sCoreWeave GB200 NVL72 (64x GB200-186GB_aarch64, TensorRT)CoreWeaveNVIDIA GB200source
7111,496.0000Tokens/sCisco UCS C880A M8 (8x NVIDIA B300-SXM-270GB, TensorRT)CiscoNVIDIA B300-SXM-270GBsource
8110,077.0000Tokens/sXA NB3I-E12ASUSTeKNVIDIA B300-SXM-270GBsource
9109,141.0000Tokens/sThinkSystem SR680a V4 (8x B300-SXM-270GB, TensorRT)LenovoNVIDIA B300-SXM-270GBsource
10107,696.0000Tokens/sG894-SD3-AAX7GigaComputingNVIDIA B300-SXM-270GBsource
11106,885.0000Tokens/sNebius B300 n1 (8x B300-SXM-270GB, TensorRT)NebiusNVIDIA B300-SXM-270GBsource
12103,961.0000Tokens/sNVIDIA DGX B300 (8x B300-SXM-270GB, TensorRT)NVIDIANVIDIA B300-SXM-270GBsource
1395,004.0000Tokens/ssmci355-ccs-aus-m09-17 (AS -4126GS-NMR-LCC)AMDAMD Instinct MI355X 288GB HBM3esource
1493,070.7000Tokens/sLLM-D v0.5.0,Openshift 4.20.12,NVIDIA 8xB200-SXM-180GBRedHatNVIDIA B200-SXM-180GBsource
1592,332.2000Tokens/sG893-ZX1-AAX4GigaComputingAMD Instinct MI355X 288GB HBM3esource
1691,846.6000Tokens/sHPE ProLiant Compute XD685 (8x AMD Instinct MI355X 288GB, vLLM)HPEAMD Instinct MI355X 288GB HBM3esource
1791,706.3000Tokens/sBM.GPU.MI355X.8ORACLEAMD Instinct MI355X 288GB HBM3esource
1891,395.5000Tokens/sPowerEdge XE9785L (8x MI355X)DellAMD Instinct MI355X 288GB HBM3esource
1985,921.2000Tokens/sNebius B200 n1 (8x B200-SXM-180GB, TensorRT)NebiusNVIDIA B200-SXM-180GBsource
2060,220.4000Tokens/sNVIDIA GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)Lambda_SITNVIDIA GB300source
2128,680.0000Tokens/sLLM-D v0.5.0,OpenShift 4.21,IBM cloud gx3d-160x1792x8h200,NVIDIA 8xH200-SXM-141GBRedHatNVIDIA H200-SXM-141GBsource
2219,040.8000Tokens/sHPE ProLiant Compute DL380a Gen12 (10x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)HPENVIDIA RTX PRO 6000 Blackwell Server Editionsource
2315,371.7000Tokens/sNebius GB300 NVL72 (4x GB300-288GB_aarch64, TensorRT)NebiusNVIDIA GB300source
2415,189.9000Tokens/sHPE ProLiant Compute DL380a Gen12 (8x NVIDIA RTX PRO 6000 Blackwell Server Edition, TensorRT)HPENVIDIA RTX PRO 6000 Blackwell Server Editionsource
257,297.0500Tokens/sHPE ProLiant DL385 Gen11 (4x NVIDIA RTX PRO 6000 Blackwell Server Edition-450W, TensorRT)HPENVIDIA RTX PRO 6000 Blackwell Server Editionsource
261,601.9100Tokens/s1-node-4x-BMG-Pro-B60-DualIntelMS-Intel Arc Pro B60 Dual 48G Turbosource
271,536.9000Tokens/s1-node-4x-BMG-B70IntelIntel(R) Arc Pro(R) B70source
28841.0380Tokens/s1-node-4x-BMG-B60IntelIntel(R) Arc Pro(R) B60source