amd-strix-halo-vllm-toolboxes/benchmarks/benchmark_results/dazipe_Qwen3-Next-80B-A3B-Instruct-GPTQ-Int4A16_tp1_throughput.json

{
    "elapsed_time": 1052.5878375879984,
    "num_requests": 100,
    "total_num_tokens": 75285,
    "requests_per_second": 0.09500394782173208,
    "tokens_per_second": 71.52372211759099
}