amd-strix-halo-vllm-toolboxes/benchmarks/benchmark_results_rocm/btbtyler09_Qwen3-Coder-30B-A3B-Instruct-gptq-8bit_tp1_throughput.json

{
    "elapsed_time": 780.1687226030044,
    "num_requests": 200,
    "total_num_tokens": 146805,
    "requests_per_second": 0.2563548040386794,
    "tokens_per_second": 188.17083503449163
}