inference/research
verifiedmedium · 0.75updated 2026-10-15

Argonne MI300X vLLM Mixtral-8x7B b64

3715.4 tok/s aggregate

Fields

id
benchmarks/argonne-mi300x-vllm-mixtral-8x7b-fp16-b64-128io-tps
Metric
tps_aggregate
Value
3,715.4
Unit
tok/s aggregate
Model
mistralai/Mixtral-8x7B-v0.1
Methodology
THIRD-PARTY MEASURED (Argonne LLM-Inference-Bench arXiv:2411.00136). vLLM, FP16/BF16, 128in/128out. throughput=batch*(in+out)/e2e latency = AGGREGATE over batch. _X_ CSV 'AMD MI300X GPU' = ALCF 8x MI300X node (xGMI). N = tiles used. 1x bs64 lat 4.41s.
Measured by
third_party
Reproducible
yes
notes
CORRECTED 2026-10-15: value re-read from All_results.csv

Notes

CORRECTED 2026-10-15: value re-read from All_results.csv

Cross-references