[Benchmark] Record model_id in bench latency/throughput --output-json (#58112)

Signed-off-by: yashasvi <yashasvi@ibm.com>

vllm/benchmarks/latency.py

@@ -169,6 +169,7 @@ def main(args: argparse.Namespace):
     # Output JSON results if specified
     if args.output_json:
         results = {
+            "model_id": args.model,
             "avg_latency": np.mean(latencies),
             "latencies": latencies.tolist(),
             "percentiles": dict(zip(percentages, percentiles.tolist())),

vllm/benchmarks/throughput.py

@@ -1141,6 +1141,7 @@ def main(args: argparse.Namespace):
     # Output JSON results if specified
     if args.output_json:
         results = {
+            "model_id": args.model,
             "elapsed_time": elapsed_time,
             "num_requests": len(requests),
             "total_num_tokens": total_num_tokens,