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

{
    "elapsed_time": 748.1414223780157,
    "num_requests": 200,
    "total_num_tokens": 146805,
    "requests_per_second": 0.2673291359329993,
    "tokens_per_second": 196.2262690032198
}