amd-strix-halo-vllm-toolboxes/benchmarks/benchmark_results_rocm/btbtyler09_Qwen3-Coder-30B-A3B-Instruct-gptq-4bit_tp1_throughput.json

{
    "elapsed_time": 621.1952276929987,
    "num_requests": 100,
    "total_num_tokens": 75285,
    "requests_per_second": 0.16097998751758127,
    "tokens_per_second": 121.19378360261106
}