From f83ea61cfe77f0b200f2ded1fda9be5ff591af2c Mon Sep 17 00:00:00 2001 From: Yashasvi Chaurasia <46622381+YashasviChaurasia@users.noreply.github.com> Date: Fri, 25 Sep 2026 22:13:37 +0530 Subject: [PATCH] [Benchmark] Record model_id in bench latency/throughput --output-json (#58112) Signed-off-by: yashasvi --- vllm/benchmarks/latency.py | 1 + vllm/benchmarks/throughput.py | 1 + 2 files changed, 2 insertions(+) diff --git a/vllm/benchmarks/latency.py b/vllm/benchmarks/latency.py index 6209f5f2d8..acb3efef15 100644 --- a/vllm/benchmarks/latency.py +++ b/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())), diff --git a/vllm/benchmarks/throughput.py b/vllm/benchmarks/throughput.py index 94b22b2956..620ce2b08f 100644 --- a/vllm/benchmarks/throughput.py +++ b/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,