Intel Arc Pro B60 24GB x4, MLPerf v6.0 Offline: gpt-oss-120b aggregate tok/s
841.038 output tok/s AGGREGATE across 4x Arc Pro B60 (Offline; ~210 tok/s per card)
Fields
id
benchmarks/b60-4x-mlperf-v6-0-gptoss-120b-offline-tokensMetric
tps_aggregate
Value
841.038
Unit
output tok/s AGGREGATE across 4x Arc Pro B60 (Offline; ~210 tok/s per card)
Model
gpt-oss-120b (FP8)
Methodology
VENDOR-MEASURED (MLPerf v6.0 closed submission, Intel). 4x Arc Pro B60 24GB GDDR6 (PCIe Gen5 x8) on Xeon 698X host, PyTorch vLLM, gpt-oss-120b, Offline. 'Tokens per second' = 841.038 verbatim from mlperf_log_summary.txt. AGGREGATE across 4 cards — ~210/card.
Measured by
vendor
Reproducible
yes
notes
gpt-oss-120b (MXFP4 MoE) — ~210 tok/s per card aggregate; the affordable 120B-class serving path.
Notes
gpt-oss-120b (MXFP4 MoE) — ~210 tok/s per card aggregate; the affordable 120B-class serving path.