amd-strix-halo-vllm-toolboxes/benchmarks/benchmark_results/dazipe_Qwen3-Next-80B-A3B-Instruct-GPTQ-Int4A16_cluster_tp2_throughput.json

{
    "elapsed_time": 764.7424081899953,
    "num_requests": 100,
    "total_num_tokens": 75285,
    "requests_per_second": 0.13076298493329488,
    "tokens_per_second": 98.44491320703105
}