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

{
    "elapsed_time": 683.9224744850071,
    "num_requests": 200,
    "total_num_tokens": 146805,
    "requests_per_second": 0.29243080533447857,
    "tokens_per_second": 214.65152188564062
}